> This is part 1 of 2 of the full documentation (pages 1–100 of 131).
> The content is paginated: fetch every part to see all of it.
> Next part: https://docs.lilypad.tech/lilypad/llms-full.txt/1
> Page index: https://docs.lilypad.tech/lilypad/llms.txt

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
This site has an agentic ask interface you may use to query the documentation dynamically by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.lilypad.tech/lilypad/index.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

---

# Lilypad Home

Lilypad is powering democratic participation in the AI Innovation Economy by pioneering new avenues for anyone to deploy, distribute and monetize AI models.

As a **full-stack, modular AI services platform**, Lilypad provides:

* A **model marketplace** for publishing and monetizing AI workloads
* **MLOps tooling** for managing and scaling AI pipelines
* A **distributed, on-demand GPU network** for high-performance inference across agent workflows, ML research and DeSci applications

Lilypad acts as a:

* **Demand engine** for decentralized compute and storage
* **Distribution layer** for training networks and agentic frameworks
* **Bridge across data, compute and AI ecosystems**, enabling straightforward integration

## Real-World Impact

In partnership with Dell, Dr. Michael Levin, and Amelie Schroder (Hugging Face's top bioML creator), Lilypad powered a protein folding AI model that discovered a novel heart medication, now progressing through wet-lab validation. We are now working with Dell to build a permissioned, licensable version of the Lilypad Network for enterprise AI adoption.

## Run AI Workloads

Lilypad makes it easy to run **containerized AI models** across a global, decentralized GPU network with no infrastructure setup required. Whether you're working with image generation models like Stable Diffusion and SDXL, text-to-video models, LLMs, or custom AI workloads, Lilypad provides the compute layer to scale inference on demand.

You can:

* **Deploy models as Lilypad Job Modules** using standard Docker containers
* **Trigger jobs via our CLI or inference API**, with full support for passing inputs and capturing outputs
* **Integrate into ML pipelines or agent workflows**, using Lilypad as a backend for large-scale AI tasks
* **Leverage the modular architecture** to compose your own workflows, fine-tune models, or chain outputs across models

If you're building an app, automating research, or experimenting with new architectures — Lilypad gives you a permissionless, cost-efficient path to run your AI anywhere.

## Provide Compute&#x20;

Join the Lilypad Network as a **Resource Provider** and earn rewards by contributing your **idle GPU or CPU power** to run real AI jobs. Whether you're a solo builder with a gaming rig or an organization managing fleets of GPUs, Lilypad lets you monetize underused compute by connecting directly to demand from developers, researchers and decentralized applications.

As a Resource Provider, you’ll:

* **Run containerized workloads** from users across the network
* **Earn rewards** for completed jobs
* **Retain full control** over your hardware with no lock-in or custodial overhead
* **Support open innovation** by enabling inference for DeSci, agent frameworks, AI art tools, and more

Lilypad’s compute layer is **permissionless**, **modular**, and designed to scale. We support **consumer GPUs**, **data center hardware**, and **containerized environments** — including Linux hosts and Docker-native setups.

## Get Started Today!

<table data-view="cards"><thead><tr><th></th><th data-hidden data-card-cover data-type="files"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>Start building with Lilypad!</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Ff55dEq9LQWQ5TwIHTg4h%2Fdevelopers.png?alt=media&amp;token=7f80c30b-d34a-4308-a51f-378f987d9c92">developers.png</a></td><td><a href="/lilypad/developer-resources/inference-api">Inference API</a></td></tr><tr><td>Become a resource provider  today!</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FBLBOSPVUlwsZ1UyPrlky%2Frps.png?alt=media&amp;token=c731fbc8-736a-4a90-aa93-494b1b4c84bd">rps.png</a></td><td><a href="/lilypad/resource-providers/rp-beta-program">RP-Beta Program</a></td></tr></tbody></table>

## Join the Community & Chat with Us

<table data-column-title-hidden data-view="cards" data-full-width="false"><thead><tr><th></th><th data-hidden></th><th data-hidden></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td>Lilypad Network</td><td></td><td></td><td><a href="https://discord.gg/tnE8SMmsxW">https://discord.gg/tnE8SMmsxW</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FQlDFte3ifp3lq6WMkPZS%2Fimage-29.webp?alt=media&amp;token=4858e186-1eb4-4fb1-8543-77f5393a7dcf">image-29.webp</a></td></tr><tr><td>Lilypad Linktree</td><td></td><td></td><td><a href="https://linktr.ee/LilypadNetwork">https://linktr.ee/LilypadNetwork</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FDi6Saj8pnHkPK7EFYrFU%2FScreenshot%202024-05-17%20at%2011.12.32%E2%80%AFAM.png?alt=media&amp;token=77432d03-9a45-4d2f-a336-d9e8866babbf">Screenshot 2024-05-17 at 11.12.32 AM.png</a></td></tr><tr><td>@Lilypad_Tech</td><td></td><td></td><td><a href="https://twitter.com/Lilypad_Tech">https://twitter.com/Lilypad_Tech</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-2428e143d3dbf118a70689e4a7befeaf691ef06b%2Ftwitter.jpeg?alt=media">twitter.jpeg</a></td></tr><tr><td>Lilypad Blog</td><td></td><td></td><td><a href="https://blog.lilypadnetwork.org/">https://blog.lilypadnetwork.org/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fv6LeXxAuc7k0dk5CqqOk%2FScreenshot%202024-05-17%20at%2011.14.42%E2%80%AFAM.png?alt=media&amp;token=50aca1ae-ed1e-49c1-a52c-9f7caa9f0444">Screenshot 2024-05-17 at 11.14.42 AM.png</a></td></tr><tr><td>@LilypadNetwork</td><td></td><td></td><td><a href="https://www.youtube.com/@lilypadnetwork">https://www.youtube.com/@lilypadnetwork</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-5533ed78e232fb1c33af71ee6f9bff1b1913ad67%2F8gzcr6RpGStvZFA2qRt4v6.jpg?alt=media">8gzcr6RpGStvZFA2qRt4v6.jpg</a></td></tr><tr><td>Lilypad-Tech Github</td><td></td><td></td><td><a href="https://github.com/Lilypad-Tech">https://github.com/Lilypad-Tech</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FYpusctzv2oKvykRDQyDA%2FScreenshot%202024-05-17%20at%2011.13.11%E2%80%AFAM.png?alt=media&amp;token=abcda1f3-11b5-440a-9b47-603db87bcfc8">Screenshot 2024-05-17 at 11.13.11 AM.png</a></td></tr><tr><td>Lilypad Luma</td><td></td><td></td><td><a href="https://lu.ma/lilypadnetwork">https://lu.ma/lilypadnetwork</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FHn7YJu4ua4JDoOxgvC40%2Fluma.webp?alt=media&amp;token=6a6cb5cb-1c18-42c6-bd26-1c4a513c53b4">luma.webp</a></td></tr></tbody></table>


# Quickstart

Get started with Lilypad

Lilypad allows you to run on-demand compute jobs on the network, enabling AI inference and high-performance workloads without managing infrastructure.

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td>Inference API</td><td>Submit inference jobs, process requests and retrieve results using Lilypad’s API</td><td><a href="/lilypad/quickstart/inference-api">Inference API</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Flq1zemExEYuHJHdlmypK%2Foutput.png?alt=media&amp;token=72074931-a932-4fb1-8c02-32c8c1379b94">output.png</a></td></tr><tr><td>CLI</td><td>Run compute jobs, manage workloads and interact with the network directly from your terminal.</td><td><a href="/lilypad/quickstart/cli">CLI</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FsvC1XZFJTIrSlcnMeBHC%2Foutput.png?alt=media&amp;token=bf602e62-f0fe-4153-9907-cbea23cf9325">output.png</a></td></tr></tbody></table>


# Inference API

Get started with the Lilypad inference API

Lilypad allows you to run on-demand compute jobs on the network using the Lilypad API, enabling AI inference and high-performance workloads without managing infrastructure.

## Overview

This guide will take you through:

* Obtaining [API keys](/lilypad/quickstart/inference-api/installation)
* [Basic usage](/lilypad/quickstart/inference-api/usage)


# Setup

Instructions for getting started with the Lilypad API

To interact with Lilypad's API, you'll first need to create an account and obtain an **API key** for authentication. This key allows you to submit jobs and retrieve results securely.

### Sign Up

Visit the [Lilypad API Dashboard](https://anura.lilypad.tech/) and sign up using an email address.&#x20;

### Generate API key

Next, log in to the dashboard and click "Create API key".

Now that you've created a new API key you can copy and securely store your API key for usage.


# Usage

Instructions on how to use the Lilypad API

Once you have your API key, you can start running AI inference jobs using Lilypad's Anura API. Below is a simple **"Hello World"** example to get started.

### Get Available Models

Before running a job, check which models are supported:

```bash
curl -X GET "https://anura-testnet.lilypad.tech/api/v1/models" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY"
```

### Run a Job

Use the API to send a simple chat completion request using one of the available models:

```bash
curl -X POST "https://anura-testnet.lilypad.tech/api/v1/chat/completions" \
  -H "Content-Type: application/json" \
  -H "Accept: text/event-stream" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "MODEL_NAME:MODEL_VERSION",
    "messages": [
      {
        "role": "system",
        "content": "You are a helpful AI assistant"
      },
      {
        "role": "user",
        "content": "What order do frogs belong to?"
      }
    ],
    "temperature": 0.6
  }'
```

Replace `"MODEL_NAME:MODEL_VERSION"` with the model you want to use from the previous step.

### Retreive Job Results

You can check the status of your job and retrieve outputs using the job ID:

```bash
curl -X GET "https://anura-testnet.lilypad.tech/api/v1/jobs/{job_id}" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY"
```

This will return details on whether the job is still processing or has completed.

For more advanced usage, including streaming completions, multi-turn conversations, and additional endpoints, check out the [inference API Documentation](https://docs.lilypad.tech/lilypad/developer-resources/inference-api).


# CLI

Get started with Lilypad

Lilypad allows you to run on-demand compute jobs on the network using the Lilypad CLI, enabling AI inference and high-performance workloads without managing infrastructure.

## Overview

This guide will take you through:

* [Installing](/lilypad/quickstart/cli/installation) the Lilypad CLI.
* [Setting up your wallet](/lilypad/quickstart/cli/setting-up-your-wallet).
* [Running a Cowsay job on Lilypad Network](/lilypad/quickstart/cli/usage), the Lilypad module equivalent to Hello, World!


# Installation

Instructions for installing the Lilypad CLI on your machine

The installation process for the Lilypad CLI involves several automated steps to configure it for your specific system. Initially, the setup script identifies your computer's architecture and operating system to ensure compatibility. It will then download the latest production build of the Lilypad CLI directly from the official GitHub repository using `curl` and `wget`.

Once the CLI tool is downloaded, the script sets the necessary permissions to make the executable file runnable. It then moves the executable to a standard location in your system's path to allow it to be run from any terminal window.

{% hint style="info" %}
When using the CLI always ensure you're running the [latest version](https://github.com/Lilypad-Tech/lilypad/releases) of Lilypad.
{% endhint %}

## Install via officially released binaries

Lilypad offers two distinct installation options to cater to different roles within the network:

* One for Lilypad users who wish to run compute jobs on the Lilypad Network.
* Another for resource providers who supply the computational resources to the Lilypad Network.

Select the appropriate installation based on your role:

{% tabs %}
{% tab title="CLI Users" %}

```bash
# Detect your machine's architecture and set it as $OSARCH
OSARCH=$(uname -m | awk '{if ($0 ~ /arm64|aarch64/) print "arm64"; else if ($0 ~ /x86_64|amd64/) print "amd64"; else print "unsupported_arch"}') && export OSARCH;
# Detect your operating system and set it as $OSNAME
OSNAME=$(uname -s | awk '{if ($1 == "Darwin") print "darwin"; else if ($1 == "Linux") print "linux"; else print "unsupported_os"}') && export OSNAME;
# Download the latest production build
curl https://api.github.com/repos/lilypad-tech/lilypad/releases/latest | grep "browser_download_url.*lilypad-$OSNAME-$OSARCH-cpu" | cut -d : -f 2,3 | tr -d \" | wget -i - -O lilypad

# Make Lilypad executable and install it
chmod +x lilypad
sudo mv lilypad /usr/local/bin/lilypad
```

{% endtab %}

{% tab title="Resource Providers" %}
{% hint style="warning" %}
The resource provider version of Lilypad is not supported on Darwin/macOS.
{% endhint %}

```bash
# Detect your machine's architecture and set it as $OSARCH
OSARCH=$(uname -m | awk '{if ($0 ~ /arm64|aarch64/) print "arm64"; else if ($0 ~ /x86_64|amd64/) print "amd64"; else print "unsupported_arch"}') && export OSARCH;
# Detect your operating system and set it as $OSNAME
OSNAME=$(uname -s | awk '{if ($1 == "Darwin") print "darwin"; else if ($1 == "Linux") print "linux"; else print "unsupported_os"}') && export OSNAME;
# Download the latest production build
curl https://api.github.com/repos/lilypad-tech/lilypad/releases/latest | grep "browser_download_url.*lilypad-$OSNAME-$OSARCH-gpu" | cut -d : -f 2,3 | tr -d \" | wget -i - -O lilypad

# Make Lilypad executable and install it
chmod +x lilypad
sudo mv lilypad /usr/local/bin/lilypad
```

{% endtab %}
{% endtabs %}

## Set WEB3\_PRIVATE\_KEY

You're required to set your private key environment variable, `WEB3_PRIVATE_KEY`, to interact securely with the network.

{% hint style="info" %}
A `WEB3_PRIVATE_KEY` can be retrieved from the Metamask account details menu. For more info, check out the [official guide from Metamask](https://support.metamask.io/configure/accounts/how-to-export-an-accounts-private-key/) on viewing a wallet's private key. **Be sure to keep your private key safe** and never share it or store it in unsecured places to prevent unauthorized access to your funds. You also need a separate private key&#x20;
{% endhint %}

```bash
export WEB3_PRIVATE_KEY=<your private key>
```

To use the Lilypad CLI, the set private key will need to hold Lilypad testnet tokens and Arbitrum Sepolia ETH. You can find those instructions in the [Setting Up Your Wallet](/lilypad/quickstart/cli/setting-up-your-wallet) documentation.

## Verify installation

To verify the installation, running `lilypad` in the terminal should display a list of available commands, indicating that Lilypad CLI is ready to use.

```bash
Lilypad: <VERSION>
Commit: <COMMIT>

Usage:
  lilypad [command]

Available Commands:
  completion        Generate the autocompletion script for the specified shell
  help              Help about any command
  jobcreator        Start the lilypad job creator service.
  mediator          Start the lilypad mediator service.
  pow-signal        Send a pow signal to smart contract.
  resource-provider Start the lilypad resource-provider service.
  run               Run a job on the Lilypad network.
  solver            Start the lilypad solver service.
  version           Get the lilypad version

Flags:
  -h, --help             help for lilypad
  -n, --network string   Sets a target network configuration (default "testnet")

Use "lilypad [command] --help" for more information about a command.
```

Thats it! You've successfully installed the Lilypad CLI on your machine! [🎉](https://emojipedia.org/party-popper)

## Uninstall Lilypad

To uninstall Lilypad, you'll need to remove the `lilypad` binary.

The following command detects where the `lilypad` binary is located and removes it.

> 🚨 Using `sudo rm -rf` can be dangerous if not used carefully. Proceed with caution.

```
sudo rm -rf $(which lilypad)
```

You can verify Lilypad has been uninstalled successfully with the following command:

```shell
which lilypad
```

If the uninstall was successful, you should see the message `lilypad not found`.


# Setting Up Your Wallet

Configure a crypto wallet to receive testnet tokens used to interact with the Lilypad Network

Both Resource Providers (GPU compute nodes) and those looking to run jobs on the Lilypad network need to set up a Metamask account in order to run jobs on Lilypad. The public key of your wallet address is how you are identified on the network, and is how you can look up the transactions you make on the [Arbitrum Sepolia blockchain explorer](https://etherscan.io/token/0xb50721bcf8d664c30412cfbc6cf7a15145234ad1). You need an account for running jobs, and a separate account for each GPU you want to set up on the network.

The wallet you use for your account must have both ETH (to run smart contracts on Ethereum) and Lilypad (LP) tokens in order to pay for jobs (or receive funds for jobs) on the network.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FBiX8YJ4pmEi1ERZ92SMH%2Feth-lp-wallet.png?alt=media&amp;token=913f822f-7c9f-4e07-b7ca-7625d4ebbe7d" alt="Metamask wallet funded with ETH Arbitrum and Lilypad tokens" width="358"><figcaption><p>A wallet funded with Arbitrum ETH and Lilypad tokens</p></figcaption></figure>

## Create a MetaMask wallet

End users of Lilypad can decide which crypto wallet they would like to use. In this guide, we advise using a [MetaMask](https://metamask.io/) crypto wallet.

* Install [MetaMask Extension](https://metamask.io/download/)

## Connect to the Arbitrum Sepolia Testnet in MetaMask.

The Lilypad Testnet (IncentiveNet) is currently running on the Arbitrum L2 network built on Ethereum.

In order to change to the Arbitrum network in the wallet, open MetaMask and click the network button in the top left of the menu bar:

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FKa5sikdV624s8IbvDUji%2Fmetamask-step-1.png?alt=media&amp;token=47aa5bbf-5d80-4e2b-ad64-d3bfd63cc70b" alt="" width="355"><figcaption></figcaption></figure>

Then select "Add network":

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F6fmBzjxKnIh073YSbe9b%2Fmetamask-step-2.png?alt=media&amp;token=0390919a-c922-4396-a0d5-a8de58d8b61b" alt="" width="352"><figcaption></figcaption></figure>

Next, select "Add a network manually":

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FcBq9iipyfjbkFCViYSXh%2Fmetamask-step-3.png?alt=media&amp;token=7ebed979-4388-4f9b-a74d-8101883679e2" alt="" width="563"><figcaption></figcaption></figure>

Input the required Arbitrum Sepolia Testnet network info, and then "Save":

{% hint style="info" %}
Network info is referenced directly from the Arbitrum Sepolia [documentation](https://docs.arbitrum.io/arbitrum-bridge/quickstart#step-2-add-the-preferred-network-to-your-wallet).
{% endhint %}

* **Network name**: Arbitrum Sepolia
* **New RPC URL**: <https://sepolia-rollup.arbitrum.io/rpc>
* **Chain ID**: 421614
* **Currency symbol**: ETH
* **Block explorer URL**: <https://sepolia.arbiscan.io> (optional)

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FtSgHdKEVCISAjuRerf4S%2Fmetamask-step-4.png?alt=media&amp;token=219a5956-f598-432c-b193-71ac4fda633a" alt="" width="375"><figcaption></figcaption></figure>

### Import the Testnet LP token

The wallet is now setup and will display an ETH (Arbitrum Sepolia) token balance. In order to also display the LP token balance, the LP token will need to be imported.

Select "Import tokens" from the three dot menu next to the network name:

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FRZvv2M8GzSUO5pxfG4tO%2Fimport%20token.png?alt=media&amp;token=8a863119-4e7a-4c8a-8d8d-ed8d3501e163" alt="" width="351"><figcaption></figcaption></figure>

Select "Custom token" and add the Lilypad token contract address and token symbol. Then "Save".

* Token contract address: `0x0352485f8a3cB6d305875FaC0C40ef01e0C06535`
* Token symbol: `LP`

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FF7VldgbaDLJ9wljK6BWv%2Fimport-token-step-2.png?alt=media&amp;token=9e5a173c-1345-4a14-98f8-98ab99a1d8fd" alt="" width="350"><figcaption></figcaption></figure>

You should now see both ETH and LP listed in the wallet (initial ETH and LP balances will be 0).

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F7qjO6VZPUuzlZnQG3wKj%2Fimport-token-step-3.png?alt=media&amp;token=3cd4d212-1e6e-421d-83ab-4d9f7b7f8a2d" alt="" width="352"><figcaption></figcaption></figure>

Now you're ready to fund the wallet with testnet LP and ETH tokens!

## Funding Your Wallet

Get Testnet Lilypad Tokens (LP) and Arbitrum Sepolia Testnet ETH

**tldr:** To obtain funds, first ensure the wallet is connected to the Arbitrum Sepolia network. then, collect LP and ETH tokens from these faucets:

* [Lilypad Testnet tokens (LP)](https://faucet-testnet.lilypad.tech/)
* [Arbitrum Sepolia ETH](https://arbitrum.faucet.dev/ArbSepolia) (3rd party faucet list)

## Faucet Guide

### Get Testnet LP tokens

Find out why you need tokens in the [FAQs](/lilypad/faq)

{% hint style="info" %}
You must be a member of [the Lilypad Discord](https://discord.gg/tnE8SMmsxW) to claim tokens
{% endhint %}

Follow these steps to successfully claim your Testnet LP tokens:&#x20;

1. Navigate to the Lilypad Testnet [faucet](https://faucet-testnet.lilypad.tech/).&#x20;
2. Authenticate with Discord.
3. Copy your MetaMask wallet address into the input.
4. Click "Request".

Testnet LP tokens will be sent to the wallet address provided to the faucet.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FX6FGac1yIekCCbKeoOq2%2Ffaucet-step-1.png?alt=media&amp;token=56c59ef5-43f9-4bda-8333-f26440994137" alt="" width="563"><figcaption></figcaption></figure>

### Get Arbitrum Sepolia Testnet ETH

Get Arbitrum Sepolia ETH from this list of [third party faucets](https://arbitrum.faucet.dev/ArbSepolia). Each faucet is designed differently, so follow the instructions provided.

{% hint style="info" %}
If you need more tokens and already have Sepolia ETH, use the [official Arbitrum bridge](https://bridge.arbitrum.io/) to transfer the tokens over to Arbitrum Sepolia.
{% endhint %}

### View tokens

With a balance of both LP and ETH, you're ready to run jobs with the Lilypad CLI!

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FGTChhjXLk5OoIRqsbulA%2Ffunded-wallet.png?alt=media&amp;token=2bf48b60-4f21-4a94-8c87-6ede0f078a0c" alt="" width="352"><figcaption></figcaption></figure>


# Usage

Running the most important Hello World on Lilypad!

Before you run a Lilypad job, make sure you have [Lilypad CLI installed](/lilypad/quickstart/cli/installation) and have set [a `WEB3_PRIVATE_KEY` env variable](/lilypad/quickstart/cli/installation#set-web3_private_key) in your environment.

{% hint style="info" %}
Your `WEB3_PRIVATE_KEY` can be retrieved from the MetaMask account details menu.  For more info, check out the [official guide from MetaMask](https://support.metamask.io/managing-my-wallet/secret-recovery-phrase-and-private-keys/how-to-export-an-accounts-private-key/) on how to get a your private key. **Be sure to keep your private key safe** and never share it or store it in unsecured places to prevent unauthorized access to your funds.
{% endhint %}

## Run Cowsay

`cowsay` is a classic CLI tool that prints messages in a speech bubble from an ASCII cow. You can run it as a job on the Lilypad network.

{% hint style="info" %}
By default, Lilypad CLI saves results to a temporary directory. To specify a custom directory for your job results, set the `DATA_DIR` environment variable before running your command: `export DATA_DIR=~/<DIRECTORY>`
{% endhint %}

Run the command:

```bash
lilypad run cowsay:v0.0.4 -i Message="moo"
```

Wait for the compute to take place and for the results to be published:

```
⠀⠀⠀⠀⠀⠀⣀⣤⣤⢠⣤⣀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⢴⣿⣿⣿⣿⢸⣿⡟⠀⠀⠀⠀⠀    ██╗     ██╗██╗  ██╗   ██╗██████╗  █████╗ ██████╗
⠀⠀⣰⣿⣦⡙⢿⣿⣿⢸⡿⠀⠀⠀⠀⢀⠀    ██║     ██║██║  ╚██╗ ██╔╝██╔══██╗██╔══██╗██╔══██╗
⠀⢰⣿⣿⣿⣿⣦⡙⣿⢸⠁⢀⣠⣴⣾⣿⡆    ██║     ██║██║   ╚████╔╝ ██████╔╝███████║██║  ██║
⠀⣛⣛⣛⣛⣛⣛⣛⠈⠀⣚⣛⣛⣛⣛⣛⣛    ██║     ██║██║    ╚██╔╝  ██╔═══╝ ██╔══██║██║  ██║
⠀⢹⣿⣿⣿⣿⠟⣡⣿⢸⣮⡻⣿⣿⣿⣿⡏    ███████╗██║███████╗██║   ██║     ██║  ██║██████╔╝
⠀⠀⢻⣿⡟⣩⣾⣿⣿⢸⣿⣿⣌⠻⣿⡟⠀    ╚══════╝╚═╝╚══════╝╚═╝   ╚═╝     ╚═╝  ╚═╝╚═════╝ v2.13.0
⠀⠀⠀⠉⢾⣿⣿⣿⣿⢸⣿⣿⣿⡷⠈⠀⠀
⠀⠀⠀⠀⠀⠈⠙⠛⠛⠘⠛⠋⠁⠀ ⠀⠀⠀   Decentralized Compute Network  https://lilypad.tech

🌟  Lilypad submitting job
2025-03-05T12:56:38-06:00 WRN ../runner/work/lilypad/lilypad/cmd/lilypad/utils.go:63 > failed to get GPU info: gpuFillInfo not implemented on darwin
2025-03-05T12:56:38-06:00 INF ../runner/work/lilypad/lilypad/pkg/web3/sdk.go:209 > Connected to arbitrum-sepolia-rpc.publicnode.com
2025-03-05T12:56:38-06:00 INF ../runner/work/lilypad/lilypad/pkg/jobcreator/run.go:27 > Public Address: 0xB86bCAe21AC95BCe7a49C057dC8d911033f8CB7c
Enumerating objects: 42, done.
Counting objects: 100% (22/22), done.
Compressing objects: 100% (4/4), done.
Total 42 (delta 18), reused 19 (delta 18), pack-reused 20 (from 1)
💌  Deal agreed. Running job...
🤔  Results submitted. Awaiting verification...
🤔  Results submitted. Awaiting verification...
✅  Results accepted. Downloading result...
🆔  Data ID: QmP2SQttNC3Hrh2xpY7bNHzV2jHq7MbfLahRC46DVzn5rG

🍂 Lilypad job completed, try 👇
    open /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q
    cat /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q/stdout
    cat /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q/stderr
```

View your results:&#x20;

```bash
cat /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q/stdout
```

```bash
~ % cat /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q/stdout
 ________________ 
< hello, lilypad >
 ---------------- 
        \   ^__^
         \  (oo)\_______
            (__)\       )\/\
                ||----w |
                ||     ||
```


# Why Lilypad

Lilypad provides distributed computational services underpinned by the [Bacalhau Project](https://www.bacalhau.org/). The network provides infrastructure for use cases like AI inference, ML training, DeSci and more. Lilypad strategically collaborates with decentralized infrastructure networks, such as Filecoin, to formulate a transparent, efficient, and accessible computational ecosystem.

Perform off-chain decentralized compute over data, with on-chain guarantees. Call this functionality directly from a smart contract, CLI, and an easy to use abstraction layer.

{% embed url="<https://www.youtube.com/watch?v=yQnB2Yxia4Y>" fullWidth="false" %}
In the video above, Ally Haire, Co-Founder of Lilypad, provides an introduction to the project and its goals.
{% endembed %}

The network is actively laying groundwork for multi-chain integration and the deployment of an incentivized testnet.

Lilypad has evolved from earlier versions (v0, v1 & v2), where the network served as a proof of concept for verifiable, decentralized compute directly from smart contracts. These earlier iterations established the groundwork for what is now a robust, scalable platform with expanded features and multichain support.

Bacalhau has been integral to Lilypad since its early versions (v0 and v1), serving as the backbone for verifiable off-chain compute. In these early iterations, Bacalhau was used for Proof of Concept projects, helping users execute decentralized compute jobs from smart contracts.

Find Lilypad on [GitHub](https://github.com/Lilypad-Tech/lilypad) or visit [the blog](https://blog.lilypadnetwork.org/).

## Objective and problem statement

Lilypad aims to mitigate the challenges predominantly associated with the accessibility of high-performance computational hardware. At present, numerous barriers impede developers and organizations from smoothly integrating projects that require high-performance computing, such as AI technologies, into their applications.

Unlike conventional centralized systems, where access to powerful compute hardware is restricted and costly, Lilypad endeavors to democratize this access. Through its verifiable, trustless, and decentralized computational network, Lilypad extends unrestricted, global access to computational power. By leveraging decentralized infrastructure networks such as Filecoin, Lilypad is strategically positioned to enhance the efficiency, transparency, and accessibility of high-performance computing hardware.

## Applications

Perform off-chain decentralized compute over data, with on-chain guarantees, and to call this functionality directly from a smart contract, CLI, API and an easy to use abstraction layer, opens the door to a multitude of possible applications including:

* Inference AI jobs
* ML training jobs
* Invoking & supporting generic ZK computations
* Cross-chain interoperability complement to bridge protocols
* Federated Learning consensus (with Bacalhau insulated jobs)
* IOT & Sensor Data integrations
* Providing a platform for Digital twins
* Supply chain tracking & analysis
* ETL & data preparation jobs

## Key features

Some of the key features of Lilypad include:

1. **Verifiable Serverless Decentralized Compute Network**: Lilypad is a decentralized compute network that aims to provide global, permissionless access to compute power. The Network orchestrates off-chain compute (a global GPU marketplace) and uses on-chain verification (Arbitrum L2 on Ethereum) to provide guarantees of compute success.
2. **Mainstream Web3 Application Support**: Lilypad is designed to enable mainstream web3 applications to use its compute network with the [API](/lilypad/quickstart/inference-api), [Lilypad CLI](https://docs.lilypad.tech/lilypad/lilypad-testnet/install-run-requirements) and [Smart Contracts](https://docs.lilypad.tech/lilypad/developer-resources/lilypad-smart-contracts). It aims to make decentralized AI compute more accessible, efficient, and transparent for developers and users.
3. **Open Compute Network**: Lilypad is an open compute network allowing users to access and run AI models/other programs in a serverless manner. Module creators and general users can access a curated set of modules or can easily create their own Lilypad module to run an AI model/other program on the network.
4. **Multichain Support**: The Lilypad Incentivized Testnet [first launched](https://blog.lilypadnetwork.org/lilypad-project-report-june-19-2024#heading-engineering-update) on Arbitrum in June 2024 with plans to go multi-chain in the near future. Supporting multiple blockchain networks will increase the scalability and interoperability of the network, allowing users to choose the blockchain that best suits their needs.
5. **Incentivized Test Net**: The [Lilypad IncentiveNet](https://lilypadnetwork.notion.site/Leap-into-Lilypad-s-IncentiveNet-9e9b12936d4340ad9417d92dab8bd9d1) is live! The IncentiveNet program provide users with[ Lilybit\_ rewards ](https://blog.lilypadnetwork.org/incentivenet-lilybit-reward-calculations)to participate in running nodes, testing, and improving the network. Learn more by checking out the [IncentiveNet Leaderboard](https://info.lilypad.tech/leaderboard).
6. **Decentralization of Mediators**: The team also aims to decentralize the mediators in the network. This means that the decision-making process and governance of the network will be distributed among multiple participants, ensuring a more decentralized and resilient system.

***

{% embed url="<https://youtu.be/9lF7omNEK-c>" %}

{% embed url="<https://youtu.be/4GgM1jY4wOw>" %}
Filecoin Unleashed \[July '23] - Lilypad v1 Unlocks internet-scale web3 applications
{% endembed %}


# Lilybit\_ Rewards

Information about the Lilybit\_ rewards program

There are multiple ways to participate in the Lilypad ecosystem and earn rewards for your contributions.

{% hint style="info" %}
Contributing to the Lilypad Network will earn Lilybit\_ credits, which will be redeemable for Lilypad (LP) mainnet tokens at the Token Generation Event (TGE).
{% endhint %}

## Developers

* [Bounty Board](https://lilypadnetwork.notion.site/bounty-board)

## Community Members

* [Community Leaderboard](https://oss.lilypad.tech/)
* [Ambassador Program](https://docs.google.com/forms/d/e/1FAIpQLSdsJHkYlVVm9YILPodExa4AgGCfW-8bHZanf9OS57QxQee3SA/viewform)

## Resource Providers

In exchange for providing compute power to Lilypad's decentralized compute network, Resource Providers (RPs) accumulate Lilybit\_ credits based on their contributions to the network.

RP incentives on Testnet will be provided via the Lilypad RP Beta Program. Any RP can submit info to join the RP beta program, although RPs with more resources will be prioritized in order to run large models on Lilypad Network.

* [Minimum hardware requirements](/lilypad/resource-providers/hardware-requirements)
* [Lilypad RP Beta Program](https://docs.google.com/forms/d/e/1FAIpQLSeF7xIHuCpwY0X44dqnl4u3weuvmtd5MkZKY0IPlGck4kHx3w/viewform)
* [Lilybit\_ Rewards](https://rp-points.lilypad.tech/)


# BUILDER-VERSE

[Builder-verse](https://blog.lilypadnetwork.org/lilypad-builder-verse) is a space for developers, labs, and builders to create powerful tools and workflows on the Lilypad Network. It’s part of our mission to accelerate open source AI through decentralized compute infrastructure and accessible tooling.

### What is it?

An open, ongoing program designed to help you build with Lilypad. There are no deadlines or fixed formats. Whether you're creating modules, building agent workflows or integrating Lilypad into a larger stack, Builder-verse gives you the space, support and rewards to grow your ideas.

### How to participate

Lilypad rewards contributors who help grow the network, whether you're building modules and agents or providing GPU compute.

You don’t need to wait for a hackathon. Submit anytime and track your impact on the [leaderboard](https://oss.lilypad.tech). Top contributors may also be eligible for additional support, amplification, and future grant opportunities.

#### Builders

* Creating and [publishing modules](/lilypad/developer-resources/module-marketplace/build-a-job-module)
* Building agents or full-stack applications that use Lilypad modules
* Driving usage. The more your module or agent is used, the more rewards you earn

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FTpomLkQzHFfDrllRwiYW%2Flp-flow.png?alt=media&amp;token=eb3bf550-ab9f-45d3-bb0a-2d189bb42dd2" alt=""><figcaption></figcaption></figure>

### What can I build?

* Your own custom AI modules
* Agent pipelines using existing modules
* Apps using Lilypad’s CLI or API
* Integrations with your existing tools or workflows

### Tools & Guides

Looking for inspiration or guidance? Check out these resources to get you started!

* [Build a module guide](https://blog.lilypadnetwork.org/lilypad-module-builder-guide) - Step-by-step guide to creating your own Lilypad module
* [`create-lilypad-module`](/lilypad/developer-resources/module-marketplace/create-lilypad-module) - CLI tool to scaffold a new module with everything you need to get started
* [Agents on Lilypad](/lilypad/use-cases-agents-and-projects/agents) - Create composable AI workflows powered by Lilypad modules

Check out our [awesome-lilypad repo](https://github.com/Lilypad-Tech/awesome-Lilypad) for a curated collection of modules, templates, and community examples!

### Get Involved

Join the Lilypad Discord and follow us on socials to stay up to date with the action!

* [Discord](https://discord.com/invite/WtHbjMP5UB)
* [Twitter/X](https://twitter.com/lilypad_tech)
* [GitHub](https://github.com/Lilypad-Tech/)
* [Read the full blog announcement](https://blog.lilypadnetwork.org/lilypad-builder-verse)
* [Explore awesome-lilypad on GitHub](https://github.com/Lilypad-Tech/awesome-lilypad)


# FAQ

Frequently Asked Questions for Lilypad Network

## 🍃 General Questions

### What is the Lilypad Network?

Lilypad is developing a serverless, distributed compute network that enables internet-scale data processing for AI, ML & other arbitrary computation from blockchains, while unleashing idle processing power & unlocking a new marketplace for compute.

Lilypad provides decentralized AI computational services. By extending unrestricted, global access to computational power, Lilypad strategically collaborates with decentralized infrastructure networks, such as Filecoin, to formulate a transparent, efficient, and accessible computational ecosystem.

Perform off-chain decentralized compute over data, with on-chain guarantees, and to call this functionality directly from a smart contract, CLI and an easy to use abstraction layer, opens the door to a multitude of possible applications.

### Lilypad Whitepaper

The Lilypad Whitepaper was planned to release by end of Q4 2024, but required further review from our team and advisors. See our [roadmap](https://lilypad.tech/#roadmap) for the current timeline.

### Roadmap

Find the full Lilypad Network [roadmap](https://lilypad.tech/#roadmap) on our website!

### What is the Bacalhau Project?

Bacalhau is a peer to peer computation network enabling compute over data jobs like GPU-enabled AI, ML, analytics, data engineering, data science, de-sci and more.\
\
With the open-source Bacalhau Project, you can streamline your existing workflows without rewriting by running Docker containers and WebAssembly (WASM) images as tasks. This architecture is also referred to as **Compute Over Data** (or CoD).

* To find out more about it, see the [Bacalhau Docs](https://docs.bacalhau.org/)

### Wait, didn’t Lilypad used to rely on determinism and optimistic reproducibility for verifiable compute?

Previously, Lilypad required deterministic jobs on the network and used optimistic reproducibility to randomly re-run jobs to ensure trust, however this method has been deprecated due to:

1. The limitation the determinism requirement placed on what jobs were available to run on the network
2. The inability to realistically be able to determine a “deterministic” job to be deterministic easily

### Has Lilypad raised VC money?

Yes, Lilypad closed our seed round of funding in March 2024.

## 🌐 Incentivized Testnet Questions

### When will the Incentivized Testnet launch?

The Lilypad Incentivized testnet launched in mid June 2024.

### How do LilyBit\_ rewards work?

Lilybit\_ rewards will be awarded to nodes for time on the network (up to a 4x multiplier) and compute power brought to the network. Rewards will be redeemable, for the Lilypad ERC20 Utility Token at Mainnet Launch, with between 5% and 10% of the total token supply (depending on IncentiveNet demand and tokenomics finalization) being allocated to this phase of the Lilypad Network.

Resource Providers (RP) can track their Lilybit\_ earnings with the [RP Leaderboard](https://info.lilypad.tech/leaderboard).

### Who can earn LilyBit\_ rewards?

Phase 1 of the Incentivized Testnet is focused on rewarding nodes on the network, referred to as Resource Providers. The earlier a provider joins the network, the more Lilybits\_ will be available.

Phases 2 and onward will provide rewards for Lilypad modules created as well as developer tools/apps (in addition to rewarding nodes).

### How do I check my Lilybit\_ rewards?

You can check your rewards by pasting your nodes wallet address into the following interfaces:

* [Grafana dashboard](https://grafana.lilypad.tech/d/adxhou3o1q8sga/rewards-per-wallets?orgId=1\&refresh=1m)
* [Lilypad leaderboard](https://info.lilypad.tech/leaderboard)
* [Open Source contributor dashboard](https://oss.lilypad.tech/)

### How does Lilypad use blockchain, and why do I need both ETH and Lilypad tokens to run a job?

On the Lilypad network, The blockchain is used for

* Payment rails
* Storing the deals transparently (on-chain guarantees about the compute)
* Storing any disputes & results

Lilypad Tokens are used to transact on the Lilypad network. They are used as payment by those who want to run jobs (to the resource providers who run them), and as collateral by resource providers.

You need ETH tokens to pay for the gas fees for smart contracts the facilitate transactions, and for records of transactions and disputes that are posted to the Ethereum blockchain.

## ⚙️ Hardware Provider Questions

### What do I need to do before I can run a Lilypad node?

The required steps before running a Lilypad node include adding the node, adding the Lilypad network information, obtaining tokens and installing the required software.

Refer to the [Run a node](/lilypad/resource-providers/run-a-node) documentation and select your preferred platform (Linux or Docker) for a detailed guide on the prerequisites.

### What are the hardware requirements to run a Lilypad node?

For more information, please visit [Hardware Requirements](/lilypad/resource-providers/hardware-requirements).

### What are the updates required for maintaining my node software?

Resource providers are expected to have the latest Lilypad versions installed on their machines. These will be included in the Docker files in each release. For more information, please refer to the [Resource Provider docs](/lilypad/resource-providers/hardware-requirements).

### How can I check the status of my Lilypad node?

Find more information check out the [Run a node](/lilypad/resource-providers/run-a-node) docs.

## 👩‍💻 Developer Questions

### What is a Lilypad Module?

View the [Modules section](/lilypad/developer-resources/module-marketplace).

### How to run a ML job on Lilypad

Follow the [quickstart guides](/lilypad/quickstart) to get started.

To build an application with Lilypad compute and modules on the backend, check out this [guide](https://blog.lilypadnetwork.org/setting-up-your-lilypad-front-end).

### How to add a ML model to run on Lilypad

A Lilypad module is a Git repository that can be used to perform various tasks using predefined templates and inputs. This ["build a job module" guide](https://docs.lilypad.tech/lilypad/developer-resources/build-a-job-module) will walk you through the process of creating a Lilypad module, including defining a JSON template, handling inputs, ensuring determinism, and other best practices.

### How to run a Lilypad node

Lilypad is an open network that allows anyone to contribute GPU computing capacity. Find instructions for running a node in our [Run a node](/lilypad/resource-providers/run-a-node) documentation.

### Node hardware specs

* Linux (latest Ubuntu LTS recommended)
* Nvidia GPU
* Nvidia drivers
* Docker
* Nvidia docker drivers

For more information on the requirements to run a Lilypad node, please refer to the [hardware requirements](/lilypad/resource-providers/hardware-requirements) documentation.

## 📖 Token Questions

### Expected TGE

Although the launch date is not finalized, the launch of Lilypad Mainnet and the TGE is scheduled for 2025.


# Join the Community

Find all of the Lilypad Network’s links below 👇

## Stay up to date & join the conversation!

<table data-view="cards"><thead><tr><th></th><th data-hidden></th><th data-hidden></th><th data-hidden data-card-cover data-type="files"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>🗨️ Join the community!</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FLcS3eFEloNkjaeo8qrmo%2Fdiscordimage.jpg?alt=media&amp;token=6bbed35e-5cd1-42af-9816-ca027ab707ad">discordimage.jpg</a></td><td><a href="https://discord.gg/tnE8SMmsxW">https://discord.gg/tnE8SMmsxW</a></td></tr><tr><td>🔗 @Lilypad_Tech</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-2428e143d3dbf118a70689e4a7befeaf691ef06b%2Ftwitter.jpeg?alt=media">twitter.jpeg</a></td><td><a href="https://twitter.com/Lilypad_Tech">https://twitter.com/Lilypad_Tech</a></td></tr><tr><td>📺 @LilypadNetwork</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-5533ed78e232fb1c33af71ee6f9bff1b1913ad67%2F8gzcr6RpGStvZFA2qRt4v6.jpg?alt=media">8gzcr6RpGStvZFA2qRt4v6.jpg</a></td><td><a href="https://www.youtube.com/@lilypadnetwork">https://www.youtube.com/@lilypadnetwork</a></td></tr><tr><td>💻 Lilypad Website</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FFxJhrwFjUjtkKvOsFQHl%2FScreenshot%202024-03-24%20231110.png?alt=media&amp;token=9d5dc509-1b4c-4203-a8af-6cdba41c9307">Screenshot 2024-03-24 231110.png</a></td><td><a href="https://lilypad.tech">https://lilypad.tech</a></td></tr><tr><td>📝 Lilypad Blog</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FLigFqGb3M14oKmzReBzM%2FScreenshot%202024-03-24%20231549.png?alt=media&amp;token=5a56ad1e-f1c2-4244-9ed9-54379f97b5b7">Screenshot 2024-03-24 231549.png</a></td><td><a href="https://blog.lilypadnetwork.org/">https://blog.lilypadnetwork.org/</a></td></tr><tr><td>🛠️ GitHub Repo</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FBNvreGMCBDqJ0mbBAgPr%2FScreenshot%202024-03-24%20232102.png?alt=media&amp;token=ff8e00c5-ce4d-40bf-9a24-875615cd35d1">Screenshot 2024-03-24 232102.png</a></td><td><a href="https://github.com/Lilypad-Tech">https://github.com/Lilypad-Tech</a></td></tr><tr><td>🌐 Lilypad Farcaster</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F18IAtDW7WRwLkoKs6wE9%2Ffarcaster.png?alt=media&amp;token=05b570fd-e071-4d0e-9464-3237a024f8b1">farcaster.png</a></td><td><a href="https://warpcast.com/lilypad-tech">https://warpcast.com/lilypad-tech</a></td></tr><tr><td>🌳 Linktree</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FJXjSttPblEP74pIdrDzy%2Flinktree.png?alt=media&amp;token=ca022d5f-4f14-4a32-afb3-37e52a47dd1c">linktree.png</a></td><td><a href="https://linktr.ee/LilypadNetwork">https://linktr.ee/LilypadNetwork</a></td></tr><tr><td>📅 Luma</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FHn7YJu4ua4JDoOxgvC40%2Fluma.webp?alt=media&amp;token=6a6cb5cb-1c18-42c6-bd26-1c4a513c53b4">luma.webp</a></td><td><a href="https://lu.ma/lilypadnetwork">https://lu.ma/lilypadnetwork</a></td></tr></tbody></table>


# Win Hacks & Contribute

Inspiration for projects and use cases that can be built on the Lilypad Network

## Hackathon & Project Ideas

### 🤝 Crossover POCs, Integrations & Plugins!

{% hint style="success" %}
Earn those double bounty dollars building out crossover POCs to Lilypad!
{% endhint %}

Lilypad is aiming to build out a full suite of decentralized AI projects, which means we want to collaborate with other projects in the ecosystem to make this vision a reality.

These include linking Lilypad compute with storage providers, data streams, databases & more.

This also means adding layers like privacy or extending usage into other frameworks like Unity or React Native projects.\
\
Some examples include:

* Using Filecoin aggregators like Banyan or Tableland Basin with Lilypad
* Extending the Chainsafe Unity SDK to add Lilypad jobs as an option
* Building Lilypad into Mona or creating a module to use on Mona
* Integrating with The Graph
* Doing a POC with data from someone like WeatherXM
* Adding privacy layers for data with Lit Protocol
* ZK Computations using zkSync for verifiable credentials and identity proofs without revealing sensitive information
* A decentralized oracle service that uses AI to fetch, verify, and deliver real-world data from tools like Chainlink to smart contracts

There are also opportunities to extend the functionality of Lilypad compute with items like:

* CI/CD pipelines
* Vector databases
* Autonomous agents
* Batching jobs with an external script
* Chaining jobs together in a pipeline\\

### 💅 Contribute a Module

Contributing a module is one of the coolest things to do for Lilypad. There are so many many options for compute and AI jobs that you can run with Lilypad, including adding new state-of-the-art OSS AI models

* Text to speech
* Text to opera (have you heard this one!)
* Finetuning models for SDXL and LLM
* Image -> cartoon (or anything else)
* Build a module that classifies images into predefined categories (e.g., cats, dogs)
* A basic recommendation system for movies, books, or products based on user preferences
* Personal finance tracker that uses AI to categorize and analyze spending patterns.
* A module that provides personalized tutoring and learning assistance using AI\\

As a quick refresher, the current module making pipeline ([see full "Build a job module" resource](/lilypad/developer-resources/module-marketplace/build-a-job-module)) looks like this:

1. Find or build a compute script (for example, [a Python script on Hugging Face like SDXL](https://huggingface.co/docs/diffusers/en/using-diffusers/sdxl))
2. Containerise the script with Docker ([see this blog for more info](https://blog.lilypadnetwork.org/how-to-build-a-custom-job-module-on-lilypad#heading-containerising-and-configuration))
3. Add a Lilypad Spec - this is simply a GitHub repo - [you can grab a template here](https://github.com/Lilypad-Tech/lilypad-module-boilerplate)!

{% hint style="warning" %}
Share your module with us by submitting a pull request to the [GitHub here](https://github.com/Lilypad-Tech/awesome-Lilypad?tab=readme-ov-file#modules)!
{% endhint %}

### ✨ DX Enhancements

Lilypad is still evolving and is currently in Beta phase. This means there are A LOT of opportunities to help build out better DX for your fellow developers!

One of the main opportunities is in the module making pipeline. See [#contribute-a-module](#contribute-a-module "mention") for a quick breakdown of the developer journey).\
\
Some of the opportunities to improve this process include:

{% hint style="info" %}
💚 Good first project (Lilypadawan)

💛 Moderate skill needed

❤️ Jedi mode!
{% endhint %}

* 💛 An application that can auto-containerise (or auto-dockerise) a given compute script
  * ❤️ BONUS IDEA: You could even then make a module to run the auto-containerising script on Lilypad after it's built! 🧙
* 💚 You could go one further and also have it auto-generate the Lilypad spec as well!
* 💛💛 Ways to make it easy for folks without GPUs at home to build & test a module
* 💚 Implement developer tooling for debugging and monitoring compute jobs, providing real-time feedback and logs
* 💛 Tools that analyze and optimize the performance of compute scripts, suggesting improvements and best practices
* 💚 Implement static analysis tools to check for code quality, style, and potential issues in compute scripts
* ❤️ Create tools to automatically audit compute scripts and modules for security vulnerabilities and compliance with best practices

### 👾 End User Projects

Build out other end-user projects across a broad range of verticals. We are very interested to hear the type of projects people are looking to build on the Lilypad Network and have released a Javascript wrapper to help folks build easily.

Here are a few basic ideas of end user projects:

* **Education:** AI-powered educational tools for personalized learning experiences
* **Music:** Developing AI-driven music composition tools, or applications for real-time music analysis and recommendation
* **Finance:** Leverage AI for risk assessment and trading strategies
* **Retail:** Building recommendation engines using AI for e-commerce platforms
* **Environment:** Developing solutions for environmental monitoring and pollution tracking
* **Security:** Tools using AI for auditing and security analysis

Projects across DeFi, DeSci, Gaming, & Metaverse, NFTs, IOT and more, as well [#crossover-pocs-integrations-and-plugins](#crossover-pocs-integrations-and-plugins "mention")!\\

### 🐐 Need more inspiration?

See the [Broken mention](broken://pages/sUrkRaYhhNwTpqSxCigL) page!

{% hint style="info" %}
Pssst... We'd love to share your projects here too! Feel free to [submit a PR](https://github.com/Lilypad-Tech/lilypad-docs)!
{% endhint %}

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FsgbPfuYh1kqyhSi0AaFP%2Fshowcase.gif?alt=media&amp;token=59d04c64-3dbc-4754-bdea-da6a843d0a24" alt=""><figcaption></figcaption></figure>

### ✨We would love to hear your feedback!✨

We hope this page inspires you to build amazing things with us! If you have any ideas, suggestions, or any questions, we'd love to hear from you! Join our community on Discord to share your thoughts, ideas, ask questions, and get the support you need from the Lilypad team. 🪷

[Join our Discord channel here](https://discord.gg/tnE8SMmsxW)!

**🌱Your feedback helps us improve and grow together!**\\


# Philosophy


# Quick Start

Get started with Lilypad

Lilypad enables you to run compute jobs on the network of the network of computers on Lilypad Testnet.

The cloud is just somebody else's computer...

<div data-full-width="true"><figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-490bbf3fa3a4a2f2ebf441f529291abd335270e7%2Fimage.png?alt=media" alt=""><figcaption></figcaption></figure></div>

## Overview

This guide will take you through:

* Setting up [MetaMask Wallet](https://metamask.io) for the Lilypad Testnet so that you can get ETH and LP tokens to fund smart contract calls on Ethereum's Arbitrum testnet, and pay for jobs on Lilypad Testnet.
* Funding your wallet with Lilypad Testnet tokens from the [faucet](http://faucet.lilypad.tech/)
* Running a Hello, (Lilysay) World! example compute job on the Lilypad Network :frog:


# Funding your wallet

Get Testnet Lilypad Tokens (LP) and Arbitrum Sepolia Testnet ETH

**tldr:** To obtain funds, first ensure the wallet is connected to the Arbitrum Sepolia network. then, collect LP and ETH tokens from these faucets:

* [Lilypad Testnet tokens (LP)](https://faucet-testnet.lilypad.tech/)
* [Arbitrum Sepolia ETH](https://arbitrum.faucet.dev/ArbSepolia) (3rd party faucet list)

## Faucet Guide

### Get Testnet LP tokens

Find out why you need tokens in the [FAQs](https://github.com/Lilypad-Tech/lilypad-docs/blob/main/lilypad/lilypad-testnet/..faqs.md)

{% hint style="info" %}
You must be a member of [the Lilypad Discord](https://discord.gg/tnE8SMmsxW) to claim tokens
{% endhint %}

Follow these steps to successfully claim your Testnet LP tokens:&#x20;

1. Navigate to the Lilypad Testnet [faucet](https://faucet-testnet.lilypad.tech/).&#x20;
2. Authenticate with Discord.
3. Copy your MetaMask wallet address into the input
4. Click "Request".

LP Testnet tokens will be sent to this wallet address.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FX6FGac1yIekCCbKeoOq2%2Ffaucet-step-1.png?alt=media&amp;token=56c59ef5-43f9-4bda-8333-f26440994137" alt="" width="563"><figcaption></figcaption></figure>

### Get Arbitrum Sepolia Testnet ETH

Get Arbitrum Sepolia ETH from this list of [third party faucets](https://arbitrum.faucet.dev/ArbSepolia). Each faucet is designed differently, so follow the instructions provided.

{% hint style="info" %}
If you need more tokens and already have Sepolia ETH, use the [official Arbitrum bridge](https://bridge.arbitrum.io/) to transfer the tokens over to Arbitrum Sepolia.
{% endhint %}

### View tokens

With a balance of both LP and ETH, you're ready to run jobs with the Lilypad CLI!

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FGTChhjXLk5OoIRqsbulA%2Ffunded-wallet.png?alt=media&amp;token=2bf48b60-4f21-4a94-8c87-6ede0f078a0c" alt="" width="352"><figcaption></figcaption></figure>


# Architecture

Lilypad Network Architecture

This page is a dynamic work in progress! We're currently working on some better diagrams!

{% hint style="success" %}
See the [Bacalhau documentation](https://docs.bacalhau.org/) for more information on how Bacalhau operates
{% endhint %}

### Lilypad Technical Architecture

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-8486df13c04c027c1e6de5bd8b8419ebf4db1c02%2Farch_lp_layers.png?alt=media" alt=""><figcaption><p>Components of the Lilypad Ecosystem: From SaaS to Smart Contracts</p></figcaption></figure>

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-6da1869f36cf20a02d0f47f72f1415bfb5625885%2Farch_lp_main.png?alt=media" alt=""><figcaption><p>Main architecture of the Lilypad Ecosystem</p></figcaption></figure>

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-2f7a42d67e7b3ed4d5c24bfb83818e3d9012395a%2Farch_lp_services.png?alt=media" alt=""><figcaption><p>Services in the Lilypad Ecosystem</p></figcaption></figure>

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FL0BWwP0IbBnA7mqiuLjz%2FSolver%20service%20(3).png?alt=media&amp;token=bbe6577c-f826-456d-8568-cf15df022573" alt=""><figcaption><p>Network layout and on-chain job flow</p></figcaption></figure>

### Based on Academic Research

The architecture of Lilypad is inspired by the research paper titled "Mechanisms for Outsourcing Computation via a Decentralized Market." The paper introduces MODiCuM, a decentralized system that allows for computational outsourcing in an open market. Just like MODiCuM, Lilypad aims to create an open market of computational resources by introducing various decentralized services like solver, resource provider, job creator, mediator, and directory services. MODiCuM's unique approach to deterring misbehavior in a decentralized environment through dedicated mediators and enforceable fines has influenced Lilypad's own design, particularly in the areas of dispute resolution and system integrity.

For an in-depth understanding, you can read the paper [here](https://arxiv.org/abs/2005.11429).

> **Abstract of the paper**:\
> Mechanisms for Outsourcing Computation via a Decentralized Market As the number of personal computing and IoT devices grows rapidly, so does the amount of computational power that is available at the edge. Since many of these devices are often idle, there is a vast amount of computational power that is currently untapped, and which could be used for outsourcing computation. Existing solutions for harnessing this power, such as volunteer computing (e.g., BOINC), are centralized platforms in which a single organization or company can control participation and pricing. By contrast, an open market of computational resources, where resource owners and resource users trade directly with each other, could lead to greater participation and more competitive pricing. To provide an open market, we introduce MODiCuM, a decentralized system for outsourcing computation. MODiCuM deters participants from misbehaving-which is a key problem in decentralized systems-by resolving disputes via dedicated mediators and by imposing enforceable fines. However, unlike other decentralized outsourcing solutions, MODiCuM minimizes computational overhead since it does not require global trust in mediation results. We provide analytical results proving that MODiCuM can deter misbehavior, and we evaluate the overhead of MODiCuM using experimental results based on an implementation of our platform.


# Inference API

Anura, Lilypad's official AI inference API

## Getting Started

Use Anura to start running AI inference job modules on Lilypad's decentralized compute network:

1. Get an API key from the [Anura website](https://anura.lilypad.tech/).

### NEW! See All Models Available (GraphQL)

You can run the following endpoints to view all Anura-supported models on the Lilypad network.\
\
These queries are available as graphQL queries [here](https://lilypad-quest-api.vercel.app/api/graphql).\
\
You can also view all available queries by opening [Apollo Server](https://studio.apollographql.com/sandbox/explorer) and putting this url in:&#x20;

```
https://lilypad-model-api.vercel.app/api/graphql
```

#### Curl Requests

Get all Models:

```bash
curl -X POST https://lilypad-model-api.vercel.app/api/graphql \
  -H "Content-Type: application/json" \
  -d '{"query": "{ allModels { id name category } }"}'
```

Response example:

```json
{
  "data": {
    "allModels": [
      {
        "id": "llama3.1:8b",
        "name": "Llama 3.1 8B",
        "category": "text-generation"
      },
      {
        "id": "sdxl-turbo",
        "name": "SDXL Turbo", 
        "category": "image-generation"
      }
    ]
  }
}
```

You can also fetch a selection of multiple model types using the following endpoint:

```bash
curl -X POST https://lilypad-model-api.vercel.app/api/graphql \
  -H "Content-Type: application/json" \
  -d '{"query": "{ chatModels { id name supportsTools } imageModels { id name maxPrompt } }"}'
```

### Get Started with Text Generation

1. Find which models we support:

```sh
curl GET "https://anura-testnet.lilypad.tech/api/v1/models" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY"
```

2. Choose a model, customize your request and fire away:

```sh
curl -X POST "https://anura-testnet.lilypad.tech/api/v1/chat/completions" \
-H "Content-Type: application/json" \
-H "Accept: text/event-stream" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
  "model": "MODEL_NAME:MODEL_VERSION",
  "messages": [{
    "role": "system",
    "content": "you are a helpful AI assistant"
  },
  {
    "role": "user",
    "content": "what order do frogs belong to?"
  }],
  "stream": true,
  "temperature": 0.6
}'
```

### Get Started with Image Generation

1. Find which models we support:

```bash
curl GET "https://anura-testnet.lilypad.tech/api/v1/image/models" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY"
```

2. Choose a model and generate your first image

```bash
curl -X POST https://anura-testnet.lilypad.tech/api/v1/image/generate \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{"prompt": "A spaceship parked on a lilypad", "model": "sdxl-turbo"}' \
  --output spaceship.png
```

## API Endpoints

### API Clients

{% hint style="info" %}
If you are using an API client such as Bruno or Postman, you can use our provided collections below.
{% endhint %}

{% tabs %}
{% tab title="Bruno" %}
{% file src="/files/HYD1csh2AQabPPnSbMul" %}
Bruno collection
{% endfile %}
{% endtab %}

{% tab title="Postman" %}
{% file src="/files/hwSUyi4UgOcakl3nrGl0" %}
Postman collection
{% endfile %}
{% endtab %}
{% endtabs %}

### Rate limits

Currently the rate limit for the api is set to 20 calls per second

### Get Available Models

To see which models are available:

```
curl -X GET "https://anura-testnet.lilypad.tech/api/v1/models" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY"
```

### Chat Completions API

{% hint style="info" %} <mark style="color:red;">\*</mark> = **Required**
{% endhint %}

#### Chat Completions

`POST /api/v1/chat/completions`&#x20;

**Note**: Due to the decentralized nature of the Lilypad Network we recommend using the streaming variant where possible at this time

This endpoint provides both a streaming interface using Server-Sent Events (SSE) and non-streaming interface for chat completions which is compliant with the OpenAI specification.  This means that you can plug and play Anura using the OpenAI SDK by simply passing in the Anura Url and API Key into your client like so:

```typescript
import OpenAI from 'openai';

const client = new OpenAI({
    baseURL: 'https://anura-testnet.lilypad.tech/api/v1',
    apiKey: process.env.ANURA_API_KEY || '',
});

const completion = await client.chat.completions.create({
  model: 'llama3.1:8b',
  messages: [
    { role: 'system', content: 'You are a helpful AI assistant.' },
    { role: 'user', content: 'Are semicolons optional in JavaScript?' },
  ],
});

return completion.choices[0].message.content;
```

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Accept: text/event-stream` (recommended for streaming)
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Request Parameters**

<table><thead><tr><th width="117">Parameter</th><th width="515.5">Description</th><th width="105.5">Type</th></tr></thead><tbody><tr><td><code>model</code><mark style="color:red;">*</mark></td><td>Model ID used to generate the response (e.g. <code>deepseek-r1:7b</code>). <strong>Required</strong>.</td><td><code>string</code></td></tr><tr><td><code>messages</code><mark style="color:red;">*</mark></td><td>A list of messages comprising the conversation so far. <strong>Required</strong>.</td><td>array</td></tr></tbody></table>

<details>

<summary>Optional Parameters and Default Values</summary>

<table><thead><tr><th width="197">Paraneter</th><th width="404">Description</th><th width="83.5">Default</th></tr></thead><tbody><tr><td><code>frequency_penalty</code></td><td>Number between <code>-2.0</code> and <code>2.0</code>. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.</td><td><code>0</code></td></tr><tr><td><code>max_tokens</code></td><td>The maximum number of tokens that can be generated in the chat completion.</td><td></td></tr><tr><td><code>presence_penalty</code></td><td>Number between <code>-2.0</code> and <code>2.0</code>. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics.</td><td><code>0</code></td></tr><tr><td><code>response_format</code></td><td>An object specifying the format that the model must output. <a href="https://platform.openai.com/docs/api-reference/introduction">Learn more</a>.</td><td></td></tr><tr><td><code>seed</code></td><td>If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same <code>seed</code> and parameters should return the same result. Determinism is not guaranteed, and you should refer to the <code>system_fingerprint</code> response parameter to monitor changes in the backend.</td><td><code>null</code></td></tr><tr><td><code>stop</code></td><td>Up to 4 sequences where the API will stop generating further tokens. The returned text will not contain the stop sequence.</td><td></td></tr><tr><td><code>stream</code></td><td>If set to true, the model response data will be streamed to the client as it is generated using <a href="https://developer.mozilla.org/en-US/docs/Web/API/Server-sent_events/Using_server-sent_events#Event_stream_format">server-sent events</a>.</td><td><code>false</code></td></tr><tr><td><code>stream_options</code></td><td>Options for streaming response. Only set this when you set <code>stream: true</code>.</td><td><code>null</code></td></tr><tr><td><code>temperature</code></td><td>What sampling temperature to use, between <code>0</code> and <code>2</code>. Higher values like <code>0.8</code> will make the output more random, while lower values like <code>0.2</code> will make it more focused and deterministic. We generally recommend altering this or <code>top_p</code> but not both.</td><td><code>1</code></td></tr><tr><td><code>tools</code></td><td><p>A list of tools the model may call. Currently, only functions are supported as a tool. Use this to provide a list of functions the model may generate JSON inputs for. A max of 128 functions are supported. At the moment only a select number models support tooling including:</p><ul><li>llama3.1:8b</li><li>qwen2.5:7b</li><li>qwen2.5-coder:7b</li><li>phi4-mini:3.8b</li><li>mistral:7b</li></ul></td><td></td></tr><tr><td><code>top_p</code></td><td><p>An alternative to sampling with <code>temperature</code>, called nucleus sampling, where the model considers the results of the tokens with <code>top_p</code> probability mass. So <code>0.1</code> means only the tokens comprising the top 10% probability mass are considered.</p><p></p><p>We generally recommend altering this or <code>temperature</code> but not both.</p></td><td><code>1</code></td></tr></tbody></table>

</details>

**Request Body (non-streaming)**

```json
{
  "model": "llama3.1:8b",
  "messages": [
    {
      "role": "system",
      "content": "you are a helpful AI assistant"
    },
    {
      "role": "user",
      "content": "write a haiku about lilypads"
    }
  ],
  "temperature": 0.6
}
```

**Response Format (non-streaming)**

The response is an OpenAI ChatCompletion Object with the following format:

```json
{
    "id": "jobId-Qmds4fif8RLVKrSKfWVGHe7fDkBwizzV5omd3kPSnc8Xdf-jobState-ResultsSubmitted",
    "object": "chat.completion",
    "created": 1742509938,
    "model": "llama3.1:8b",
    "system_fingerprint": "",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "Lily pads dance\nOn the water's gentle lap\nSerene beauty"
            },
            "finish_reason": "stop"
        }
    ],
    "usage": {
        "prompt_tokens": 2048,
        "completion_tokens": 184,
        "total_tokens": 2232
    }
}
```

**Response Codes**

* `200 OK`: Request successful, stream begins
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `404 Not Found`: Requested model not found
* `500 Internal Server Error`: Server error processing request

**Response Object Fields**

The response data contains the following fields:

<table><thead><tr><th width="383">Field</th><th>Description</th></tr></thead><tbody><tr><td><code>id</code></td><td>A unique identifier for the chat completion</td></tr><tr><td><code>object</code></td><td>The object type</td></tr><tr><td><code>created</code></td><td>Timestamp when the response was created</td></tr><tr><td><code>model</code></td><td>The model used for generation</td></tr><tr><td><code>choices</code></td><td>The array containing the assistant's response</td></tr><tr><td><code>choices[0].message.role</code></td><td>Always "assistant" for responses</td></tr><tr><td><code>choices[0].message.content</code></td><td>The generated text content</td></tr><tr><td><code>choices[0].message.tool_calls</code></td><td>The array containing the corresponding tool response objects (this is only applicable if you make a tool request)</td></tr><tr><td><code>choices[0].finish_reason</code></td><td>Reason for completion (e.g., "stop", "length")</td></tr><tr><td><code>usage.prompt_tokens</code></td><td>The number of tokens used in the prompt</td></tr><tr><td><code>usage.completion_tokens</code></td><td>The number of tokens in the generated completion</td></tr><tr><td><code>usage.total_tokens</code></td><td>The sum of the prompt_tokens and the completion_tokens</td></tr></tbody></table>

**Request Body (streaming)**

```json
{
  "model": "llama3.1:8b",
  "messages": [
    {
      "role": "system",
      "content": "you are a helpful AI assistant"
    },
    {
      "role": "user",
      "content": "write a haiku about lilypads"
    }
  ],
  "stream": true,
  "temperature": 0.6
```

**Response Format (streaming)**

The response is a stream of Server-Sent Events (SSE) with chunked OpenAI ChatCompletion objects with the following format:

**Initial response**:

```json
data: {
    "id": "jobId-QmZXDGS7m8VuJrURqsKvByGKHCM749NMVFmEA2hH2DtDWs-jobState-DealNegotiating",
    "object": "chat.completion.chunk",
    "created": 1742425132,
    "model": "llama3.1:8b",
    "system_fingerprint": "",
    "choices": [
        {
            "index": 0,
            "delta": {
                "role": "assistant",
                "content": null
            },
            "finish_reason": null
        }
    ],
    "usage": {
        "prompt_tokens": 0,
        "completion_tokens": 0,
        "total_tokens": 0
    }
}
```

**Processing updates**:

```json
data: {
    "id": "jobId-QmZXDGS7m8VuJrURqsKvByGKHCM749NMVFmEA2hH2DtDWs-jobState-DealAgreed",
    "object": "chat.completion.chunk",
    "created": 1742425135,
    "model": "llama3.1:8b",
    "system_fingerprint": "",
    "choices": [
        {
            "index": 0,
            "delta": {
                "role": "assistant",
                "content": null
            },
            "finish_reason": null
        }
    ],
    "usage": {
        "prompt_tokens": 0,
        "completion_tokens": 0,
        "total_tokens": 0
    }
}
```

**Content delivery**:

```json
data: {
    "id": "jobId-QmZXDGS7m8VuJrURqsKvByGKHCM749NMVFmEA2hH2DtDWs-jobState-ResultsSubmitted",
    "object": "chat.completion.chunk",
    "created": 1742425147,
    "model": "llama3.1:8b",
    "system_fingerprint": "",
    "choices": [
        {
            "index": 0,
            "delta": {
                "role": "assistant",
                "content": "Lily pads dance\nOn the water's gentle lap\nSerene beauty"
            },
            "finish_reason": "stop"
        }
    ],
    "usage": {
        "prompt_tokens": 2048,
        "completion_tokens": 456,
        "total_tokens": 2504
    }
}
```

**Completion marker**:

```
data: [DONE]
```

**Response Codes**

* `200 OK`: Request successful, stream begins
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `404 Not Found`: Requested model not found
* `500 Internal Server Error`: Server error processing request

**Response Object Fields**

The delta event data contains the following fields:

<table><thead><tr><th width="383">Field</th><th>Description</th></tr></thead><tbody><tr><td><code>id</code></td><td>A unique identifier for the chat completion</td></tr><tr><td><code>object</code></td><td>The object type</td></tr><tr><td><code>created</code></td><td>Timestamp when the response was created</td></tr><tr><td><code>model</code></td><td>The model used for generation</td></tr><tr><td><code>choices</code></td><td>The array containing the assistant's response</td></tr><tr><td><code>choices[0].delta.role</code></td><td>Always "assistant" for responses</td></tr><tr><td><code>choices[0].delta.content</code></td><td>The generated text content</td></tr><tr><td><code>choices[0].delta.tool_calls</code></td><td>The array containing the corresponding tool response objects (this is only applicable if you make a tool request)</td></tr><tr><td><code>choices[0].finish_reason</code></td><td>Reason for completion (e.g., "stop", "length")</td></tr><tr><td><code>usage.prompt_tokens</code></td><td>The number of tokens used in the prompt</td></tr><tr><td><code>usage.completion_tokens</code></td><td>The number of tokens in the generated completion</td></tr><tr><td><code>usage.total_tokens</code></td><td>The sum of the prompt_tokens and the completion_tokens</td></tr></tbody></table>

**Conversation Context**

The API supports multi-turn conversations by including previous messages in the request:

```json
{
    "model": "llama2:7b",
    "messages": [
        {
            "role": "user",
            "content": "write a haiku about lilypads"
        },
        {
            "role": "assistant",
            "content": "Lily pads dance\nOn the water's gentle lap\nSerene beauty"
        },
        {
            "role": "user",
            "content": "Now write one about frogs"
        }
    ],
    "temperature": 0.6
}
```

This allows for contextual follow-up questions and maintaining conversation history.

#### Tooling calls

The Anura chat completions endpoint supports requests with tooling allowing for function calling through many popular AI frameworks and sdks. &#x20;

At the moment only a select number models support tooling including:

* llama3.1:8b
* qwen2.5:7b
* qwen2.5-coder:7b
* phi4-mini:3.8b
* mistral:7b

Below is a sample request and response

**Request:**

```json
{
    "model": "mistral:7b",
    "messages": [
        {
            "role": "system",
            "content": "you are a helpful AI assistant"
        },
        {
            "role": "user",
            "content": "What's the weather in Tokyo?"
        }
    ],
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "get_current_weather",
                "description": "Get the current weather for a city",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "city": {
                            "type": "string",
                            "description": "The name of the city"
                        }
                    },
                    "required": [
                        "city"
                    ]
                }
            }
        }
    ],
    "temperature": 0.6,
    "stream": false
}
```

**Response:**

```json
{
    "id": "jobId-QmTm3E4oEu4TYp1FLykHdnrkPyX6cLz2UUYS45YrmrzqdN-jobState-ResultsSubmitted",
    "object": "chat.completion",
    "created": 1742790608,
    "model": "mistral:7b",
    "system_fingerprint": "",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "",
                "tool_calls": [
                    {
                        "id": "call_syyia0kt",
                        "index": 0,
                        "type": "function",
                        "function": {
                            "name": "get_current_weather",
                            "arguments": "{\"city\":\"Tokyo\"}"
                        }
                    }
                ]
            },
            "finish_reason": "tool_calls"
        }
    ],
    "usage": {
        "prompt_tokens": 88,
        "completion_tokens": 22,
        "total_tokens": 110
    }
}
```

#### Vision Support

The chat completions API also supports vision requests allowing for image-to-text search against a base64 encoded image. This will allow you to make a query against an image asking a LLM what the image is or about particular details around it. Currently vision is only supported via the following models (more coming soon):

* llava:7b
* gemma3:4b

Additionally, the vision capability is limited by the following constraints:

* Images must only be base64 encoded (you cannot pass a link to an image at this time)
* Maximum image size is 512px x 512px
* Support for JPEG or PNG format

**Request:**

```json
{
    "model": "llava:7b",
    "messages": [
        {
            "role": "system",
            "content": "you are a helpful AI assistant"
        },
        {
            "role": "user",
            "content": [
                {
                    "type": "text",
                    "text": "What's in this image?"
                },
                {
                    "type": "image_url",
                    "image_url": "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAMkAAACqCAYAAAAHpmvxAAAKpmlDQ1BJQ0MgUHJvZmlsZQAASImVlgdQU+kWgP97bzoJLXQpoTdBikAAKSG00KWDqIQkQCghBIKKDZHFFVwLIiJgQ1eago0ia0EsWFgUFEVFF2RRUNfFgg2Vd4Eh7O6b9968M3Nyvjk5/yn/3H/mAEAhs4XCFFgWgFRBpijYy40WGRVNw48CDECAIqACOzYnQ8gICvIDqMzav8uHewCasnfMpnL9+///VeS4vAwOAFAQynHcDE4qyqdRHeEIRZkAIBWoX3dFpnCK21BWEKENonx3ihNmeGSK42b463RMaDATAAw6FYHMZosSACCro35aFicBzUNehLKFgMsXoDzVr3NqahoX5aMoG6ExQpSn8tPj/pIn4W854yQ52ewECc/MMi0Ed36GMIW96v+8jv8tqSni2RoGqJITRd7BqEX7gvqS03wlLIgLCJxlPnc6fpoTxd5hs8zJYEbPMpft7is5mxLgN8vxfE+WJE8mK3SWeRkeIbMsSguW1IoXMRmzzBbN1RUnh0n8iTyWJH92YmjELGfxwwNmOSM5xHcuhinxi8TBkv55Ai+3ubqektlTM/4yL58lOZuZGOotmZ091z9PwJjLmREp6Y3Lc/eYiwmTxAsz3SS1hClBknheipfEn5EVIjmbiX6Qc2eDJHeYxPYJmmUQBLyANbACYYAJ3IF/Jm9l5tQQzDThKhE/ITGTxkBfF4/GEnDM59OsLKxsAJh6qzOfwru+6TcIKRHmfFufAuByAHUmzfmYWgCcQO+A9GbOp9cLgIwKAFd7OGJR1owPM/WDBSQgAxSAKtAEusAImKHd2QJH4Ao8gA8IBKEgCiwDHJAIUoEIrABrwAaQDwrBdrALlIH94BCoBsfASdAMzoKL4Cq4CW6DXvAIDIBh8BKMgQ9gAoIgPESBqJAqpAXpQ6aQFUSHnCEPyA8KhqKgWCgBEkBiaA20ESqEiqAy6CBUA52AzkAXoetQN/QAGoRGobfQFxiBybACrAEbwAtgOsyAfeFQeCmcAKfD2XAevBUuhSvho3ATfBG+CffCA/BLeBwBiBSihGgjZggdYSKBSDQSj4iQdUgBUoJUIvVIK9KB3EEGkFfIZwwOQ8XQMGYYR4w3JgzDwaRj1mG2YMow1ZgmzGXMHcwgZgzzHUvBqmNNsQ5YFjYSm4Bdgc3HlmCPYBuxV7C92GHsBxwOp4QzxNnhvHFRuCTcatwW3F5cA64N140bwo3j8XhVvCneCR+IZ+Mz8fn4Pfij+Av4Hvww/hNBiqBFsCJ4EqIJAkIuoYRQSzhP6CE8J0wQZYn6RAdiIJFLXEXcRjxMbCXeIg4TJ0hyJEOSEymUlETaQCol1ZOukPpJ76SkpHSk7KUWS/GlcqRKpY5LXZMalPpMliebkJnkGLKYvJVcRW4jPyC/o1AoBhRXSjQlk7KVUkO5RHlC+SRNlTaXZklzpddLl0s3SfdIv5YhyujLMGSWyWTLlMickrkl80qWKGsgy5Rly66TLZc9I3tfdlyOKmcpFyiXKrdFrlbuutyIPF7eQN5DniufJ39I/pL8EBWh6lKZVA51I/Uw9Qp1WAGnYKjAUkhSKFQ4ptClMKYor7hQMVxxpWK54jnFASVEyUCJpZSitE3ppNI9pS/KGsoMZZ7yZuV65R7ljyrzVFxVeCoFKg0qvSpfVGmqHqrJqjtUm1Ufq2HUTNQWq61Q26d2Re3VPIV5jvM48wrmnZz3UB1WN1EPVl+tfki9U31cQ1PDS0OosUfjksYrTSVNV80kzWLN85qjWlQtZy2+VrHWBa0XNEUag5ZCK6Vdpo1pq2t7a4u1D2p3aU/oGOqE6eTqNOg81iXp0nXjdYt123XH9LT0/PXW6NXpPdQn6tP1E/V363fofzQwNIgw2GTQbDBiqGLIMsw2rDPsN6IYuRilG1Ua3TXGGdONk433Gt82gU1sTBJNyk1umcKmtqZ8072m3fOx8+3nC+ZXzr9vRjZjmGWZ1ZkNmiuZ+5nnmjebv16gtyB6wY4FHQu+W9hYpFgctnhkKW/pY5lr2Wr51srEimNVbnXXmmLtab3eusX6zULThbyF+xb22VBt/G022bTbfLO1sxXZ1tuO2unZxdpV2N2nK9CD6Fvo1+yx9m726+3P2n92sHXIdDjp8KejmWOyY63jyCLDRbxFhxcNOek4sZ0OOg0405xjnQ84D7hou7BdKl2euuq6cl2PuD5nGDOSGEcZr90s3ERujW4fmQ7Mtcw2d8Tdy73AvctD3iPMo8zjiaeOZ4JnneeYl43Xaq82b6y3r/cO7/ssDRaHVcMa87HzWetz2ZfsG+Jb5vvUz8RP5NfqD/v7+O/07w/QDxAENAeCQFbgzsDHQYZB6UG/LMYtDlpcvvhZsGXwmuCOEGrI8pDakA+hbqHbQh+FGYWJw9rDZcJjwmvCP0a4RxRFDEQuiFwbeTNKLYof1RKNjw6PPhI9vsRjya4lwzE2Mfkx95YaLl259PoytWUpy84tl1nOXn4qFhsbEVsb+5UdyK5kj8ex4irixjhMzm7OS64rt5g7ynPiFfGexzvFF8WPJDgl7EwYTXRJLEl8xWfyy/hvkryT9id9TA5MrkqeTIlIaUglpMamnhHIC5IFl9M001amdQtNhfnCgXSH9F3pYyJf0ZEMKGNpRkumAroUdYqNxD+IB7Ocs8qzPq0IX3FqpdxKwcrOVSarNq96nu2Z/fNqzGrO6vY12ms2rBlcy1h7cB20Lm5d+3rd9Xnrh3O8cqo3kDYkb/g11yK3KPf9xoiNrXkaeTl5Qz94/VCXL50vyr+/yXHT/h8xP/J/7NpsvXnP5u8F3IIbhRaFJYVft3C23PjJ8qfSnya3xm/t2ma7bd923HbB9ns7XHZUF8kVZRcN7fTf2VRMKy4ofr9r+a7rJQtL9u8m7RbvHij1K23Zo7dn+56vZYllveVu5Q0V6hWbKz7u5e7t2ee6r36/xv7C/V8O8A/0HfQ62FRpUFlyCHco69Czw+GHO36m/1xzRO1I4ZFvVYKqgerg6ss1djU1teq12+rgOnHd6NGYo7ePuR9rqTerP9ig1FB4HBwXH39xIvbEvZO+J9tP0U/Vn9Y/XdFIbSxogppWNY01JzYPtES1dJ/xOdPe6tja+Iv5L1Vntc+Wn1M8t+086Xze+ckL2RfG24Rtry4mXBxqX97+6FLkpbuXF1/uuuJ75dpVz6uXOhgdF645XTt73eH6mRv0G803bW82ddp0Nv5q82tjl21X0y27Wy237W+3di/qPt/j0nPxjvudq3dZd2/2BvR23wu713c/5v5AH7dv5EHKgzcPsx5OPMrpx/YXPJZ9XPJE/Unlb8a/NQzYDpwbdB/sfBry9NEQZ+jl7xm/fx3Oe0Z5VvJc63nNiNXI2VHP0dsvlrwYfil8OfEq/w+5PypeG70+/afrn51jkWPDb0RvJt9ueaf6rur9wvft40HjTz6kfpj4WPBJ9VP1Z/rnji8RX55PrPiK/1r6zfhb63ff7/2TqZOTQraIPb0KIKjC8fEAvK0CgBIFAPU2uj8smdmlpwWa2f+nCfwnntm3p8UWgAbUTK1lzjkANKKq1waAtCsAQaiGugLY2lqis3vv9I4+JbhTAFi9RTcH6Onx/hzwD5nZ3//S9z8tkGT9m/0XpV8Ef9hUGu0AAABWZVhJZk1NACoAAAAIAAGHaQAEAAAAAQAAABoAAAAAAAOShgAHAAAAEgAAAESgAgAEAAAAAQAAAMmgAwAEAAAAAQAAAKoAAAAAQVNDSUkAAABTY3JlZW5zaG90bt216gAAAdZpVFh0WE1MOmNvbS5hZG9iZS54bXAAAAAAADx4OnhtcG1ldGEgeG1sbnM6eD0iYWRvYmU6bnM6bWV0YS8iIHg6eG1wdGs9IlhNUCBDb3JlIDYuMC4wIj4KICAgPHJkZjpSREYgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj4KICAgICAgPHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9IiIKICAgICAgICAgICAgeG1sbnM6ZXhpZj0iaHR0cDovL25zLmFkb2JlLmNvbS9leGlmLzEuMC8iPgogICAgICAgICA8ZXhpZjpQaXhlbFlEaW1lbnNpb24+MTcwPC9leGlmOlBpeGVsWURpbWVuc2lvbj4KICAgICAgICAgPGV4aWY6UGl4ZWxYRGltZW5zaW9uPjIwMTwvZXhpZjpQaXhlbFhEaW1lbnNpb24+CiAgICAgICAgIDxleGlmOlVzZXJDb21tZW50PlNjcmVlbnNob3Q8L2V4aWY6VXNlckNvbW1lbnQ+CiAgICAgIDwvcmRmOkRlc2NyaXB0aW9uPgogICA8L3JkZjpSREY+CjwveDp4bXBtZXRhPgr55CqOAABAAElEQVR4Aey92ZNk2XnYd6oys7Iya9+7et+mp3tmMBgMtsFKiSQomhaDlkzKlkIvluwI/Rn2ix8c9qNfFOFwkGGHFsoiKZkUSVAkMQRAALPv09P7vtS+ZlVlVpV/v+/cW10zmHHIBDDdjfDtvnVv3uXcc77z7d93zuk69/Tnd9P/vz2aEOjqSsndrbs7Trv83V2N611cS5Vq6qpwr1KJZ7qr/PY6e7W3kd/lnZ3tDu938Wot1Zp9qTE8lCy5vd5K6/Pz3G+zb6e0s5N22dPubtrpdNJuXNvlWJ7n+z7Ttcs5z5XPx8d89+dsA9qP+VYi0d+0GXTyI7Xtb09xHoTBeSYQjgVBlETSXYVAuNbN7r3unh7Oa2n05Mm0DZJ3QPb1e3chnkqqDw6kkSPHUv/4OESwnZZu3kpbGxtpe2szdUkUO9t8B6KDyDqt9SASCWWnXUDJOpV13KUuEEpsBXHseu9Rg+lP2MGPP5HsB0DZefuvfdz5o96J+9oRhMDvUjpwEgTRVUgMjyE9JBwJBElRHxxMtf6B9Nn/8r9Ka8uraene/XT5L/8k9fT1pdGjx9Nnfu3XwPOetHp/Nq3NLKcEUYHZCCXQAaTv6e9PNZ5dvXU7iGd7ayt1U4ddCCckC8TF41mC8KZVktWERBHe1v9Rh7H1/E/cHi8i2Yc80T5/77tWctz/17bbefs6UHVhb/vIvb3rf9OTfXXbK+Ljru3dtDn72sR5EEcqCKNQo7prqFQSiZKDY6Xemyo99VTta6YeVKmJM+fSxOlzqTk1nerj26n/wFTqbKymofGJNMC+3e5O9y9+kGYuXUzvf+/F1F3bTfVmI/Whgk0//WxW5yj7zvtXUr2BejbYSJ3V1T0i2elC4lDnLqRHVsesN1ckoo/Cc1/bHtfTx4NI9iPWvvM9hAKv6KZAqBLl497H9cqOT6BHFwQRxRUvle9+3Gt/o2v76lq+/2P1Kurut723d993ZdFeA/m6sD2UFJ7vEYkEAnEMHT2eeoeHU31oGGnQk8ZOnE4jXKvU6txPqVrrSVOnz6Y+7tcbvamztZuW7txLC7fvpLWFxdQ3OoAE6uH9UYjrybS2uJxWF5fSxFneGR4E93fT7LvvpPbaatgp1IpNOyWrWruqXNbTyxwDttbfbT/R5CuP3d9Hn0hKYAva8lykKc67ukUmO4aj1/c9Vz7jpUwUnNDhdjBvFB3oMd8PzujDH7ft7+yyHh/3HNf2vls85xf3ru0/L/CIu3E/PuE7xXvaELYJXaeQHNko71KS1GoQRC1Vm800+dTTaeT4qdQ3MYkd0U49jWaqQjy7bZF4N1SlyWMnMcwzYlerO6m1tJLW5hdDPetG9ar19qfm6FQQ2Mbb76KKXUlPfutX08DEMEVsp9V7d0Kd6my0kqVKKLtd2CrBbApiwZ6JPuDaz5Pq9WgTSYmMJeJwFNkC4eSykoRIxLXuEpnier5VIhu/MDAzMcjZgvOJuXudCdF43r1D5wfF4OSRkPZtvl9u1ofNKyXyF5e84q3YwoiVAPgVfx/cyvX2qVxIVDW3Kz/thS4M8pAiEr/qlTvn3UgG7ZDe4ZE0fvoMnqqJVOtpQv/dqX8Izq/Hi4KzIb5DW9ij/nq5MOQ3NtPh555NE0+cSmd/6Vtpt9NOdVS15vhowGbsOIY9BLeL8d9G6uy0t1HdjlMGcF5eSu3VlbRD/Xa7VbtsgDDlyN4FEwpb3v7QmN8PN9v7GG6PJpE8wDixMCOiQPe6u+eiHgQR7s/yGtfDM8MtEbREYPul7MToSHXp6FRuyP2KDqbnfTJ2dewgLH792EbZgdvc2P+NqJvXrFu57dWjuFbe2rvug1z0enFN8lRCBlFwzfZ2IznCBtE451zDugGRjBw+mpoDQ6hUdQjcf7wHlnZB5OHSDdzNCGy5bQhkZXY2bbU2UqcNknOrik2ju1j1bGt9M213QG6kmFKjs4TnC69Xc3Qs7Wytpw3czTt4wrRFhKHEKhGGbWJhnFtlN379XGyPHpGUEPbIXnJXOWiGvgiU1ZC4V+jp8aznviex0D3xDkc7M0uSAllKDmenbvOkxzgvOZ/dm1WIT+ply4+trGN8kStR7/JWUf/8YFF/n8nP7RFYvJNL9NreDgJGW7wGYYSEKIikd2gEbn8gTZ18EuKpxXMVYQTyRtUpbqewGYoPBhy2WmtpBrfvNh6q7XY7ba6spJHpqVRt9/B7O22ub4DnqFHsreXFtD43lzqbm2ns6GFNItSy3rS1tBjvyqqUUF18M4iF36FmIU32VFfb5iZ8H9Pt0SMSAVkgTalqyAEDWUQCCUBisMc8L1SQeNbrXvO5snN41/5B0Ygu2oXDGgvwoh0aneoDnEfwrCAgnTXcjHf848+iuznfh/xeFImLOvts+e391/afx/3yHZ+3vkXppY0VErEMFIZUgUhsawQOq2n67DNpEO/VFgjcWroH119PO0gJJUC1B3uFPeIm/Ja4akiL1upyWsAdfP31N1HP6uy11NfsT4Ojo8RWeoNItrFjjK0Io0qtl+t4ttZa6ZV/+2+DWKo91TR0YDKkzc7mhrXPMLQFwiDg28kwsV3COR6Kv549dtujRSQFogWwC2QPhPS8JACJw/NyDyKphJdn4NgxVIEtvDeoDK0WQbOjqa9/MPWCEJffeYtOhkvKadtb9F1WuUJtkDDYIxZQEInElJWq3KdlF0sT+Zwz/1tnN1U/T/OfB/f23S/oYE/C+WxJPLmcB78tL4KEBRyqGOP1AeIfGuX1JnGQQRhEDQTVMO9OKzOz6RZt3CIAKKH5blZFZRrABzh1gE1rdT0t3r2fenpxGSOdGqhZa6sLqG7jqX/qINJpOsoMZoKhr/esp0+jfixtLC0E8TUJRNZ6KmkLSbNx/37aqUAUwFXmtMu3Q2213WXbZUKP8fZoEEkBzJAAAlPkAeD8CSQqCSLkPcTxIf0cRNE9avBs6jOfTVura+zLaX12Jh35zHNpeGIiNeuNNLu4mNbo5O32ZuosL0MkdBwEofQAK7KOze+SaD5qjxSkEF0tHpSEUiJ3IIR4Yb1jsw3FW4Ew+Xqc0i7f910RWAOdX/F8fp/zEgaU4bX+sfHUNz6VegdHIJah1AvBVOvYITxnxHzhzu10/vt/hUt3lrIhcCROODOEYdQjE3EVeFUgDjk8JadKFyjw/ltpYPJgOnDuMxj0zYCnMNc20aXcg7Tp5ZudrQ1+G7UfxG08lNbu3YNQlnEQEI6XWVmm7WfHIca58CwIxIY/psTyaBCJSAUQA3EAcCAenavqINBDpcLtGeqDncEehKIqgcpQJYDWN3kgHfvaN3geQxKOuUEaxpGDk2l1YSVdfP9iOvq1v4O+vZqWZ26ne2++nCoiH19sY8Du0Pk76OclgUg0SpwCxTNBWcdiUxWyvm4ZATn3WvE71z8juoVkgshIWraRF0J9qjT6IsItx3bPREJBINQ2yF9Ks+e++Yup2j+UfN72V/h+1ABJ8v6Lf5luIEXa7Y3UO0S0HLuhyt6LtDECr7oVRjbv9A8PpBPPPJEuvvJ6mr12Pd18+z2au53uXXg3XX7pxXSaAOPkmafZnwpDfpsyN9dWcAnfxahfpdq76dYbb6Zzv/Kr5ID1U8+dNP/eu6kDg9EOCdVRQpEgIBThJDyCKQmgx3B7+EQCAB8gHIgkF6L7g8OGSsW1iDAjMYrj+FPPAOqu8LoMHZyGw3WIOoNgdFK1BnckdtA8MJ3Wt3bSZndPaqBCtPHUtEE6JUcv8YAqqRjGGVrECjZWSPBDTanswhFJ5AvVS6nyCZwv5zYFiu550ayvhBt5U4WLNscKMpJUkWbVujlV1bRw80Ygb0/fQBo8cCTqXpEJ8P42UlBpuLmyRLyjNw2OESVHGtaHxo0KRrt34fBaSxsry+n+lUtp5saVgM2Rz30RA/tI6h8dTs3h/rDhM5J2g/DVVKXKbVy46zN30ursQtqEQdQJFnY2t6LNuxjz9z94M20sz6fFm1fS8Rd+IeAlI9Go31pbgzAaaXD6YGpvtiHqZho++URavnEdpoKnrFC5glAgXh3qmVI4PMbbo0EkAHBPvQjOk4kldGqlhsQiEsltMUAHDx0JlUAde/TEieg8EVLeypOpBiLW+4bS4uxc2sQQrcLx1pfvoK+vBULUGv2hqlSquDxbfBvjtwsE6SLOEJwcXSEb9FkK2L9BtNYNIhaZJVI3JY7181q4UUMiQIAcTSD0Od+pQCQiva+tEtGuYhTXB4bTwNTRUI2i/l0G+Uj/6OxiLG+moZHJNDBGgI/2KkF8RnerRNzeakdA8PbF86Fu9YC840bajxxOQ1OoZiODMAWI3mpSb6PuNcrfXFhIl27dpP0SbSONwGRWZufDNby9u0GeFwmPMIzVuftp/NTZaHdWmXD90p5tVCttFBmIUrs5MhGSvE3ayp4UKRhfSJDiPBjhJzAd4fgobw+fSIBOqZ4IKJFRYIdKhSqlbhzcGaRrEOAaPHIiNScOoZ+Pp8ZQTvfWCK1DPGAPrBIER0WQWOZv3UoryytpmyDY4s3LcEi46JIcuo/Oxs2JvtxA167UGmm3iSG6tpxqNb5HJLlLiSKCg2SqSI3RSWITGM7NgTR88DAIs4MEa6el+zOpiYrTQE8fIK1DRJJYNZzbqDEmAtie1VnUFdI6tiHI41/+BeogoeI94h56TGotzKVbr7+c7l28FERXBxG/8MJvQnzklSD12qiDocJQGe2QOxfeT/euXEgXXvrr9NS3fgkJchx4DKbX/v0fhsQUjqMHD4bBrldss7UJYZGoiK0yOHUsffWFr0E4VVJQFtJL/9fvYfjPpPXFubR69x7Et5DWqc+rv/vb2ChfSsOHTwDvgyFNUlc7ou9DwgD4rS+uwcCwjahnV1drry+j3TQ9ACgQy/PHkFAeHpHsB5z2Qej5IIwEQgdHtJnOzfZItjvqEEXfxHggkZirB6bqMzJXkYit2yAgN7u5t7WO2mK+EWK/2ttMDThfpd6PHUJqxTYeLvaVmbl4j5eQPoOpp4LkoT71GhIDwhQBKgTuRkn7kIPqXl2D228TiOuqddLBpybAYVQZiLlK+dZpG+LooKvXwiCnnkibVaSazKCCq7WXdPU6dlQvnqUm5bVmIep6JU0eOpTGJye5P4wadiB/vwumQZkqL0GtSJGZW9fSpVd/kFaJV0w+eTZVSClpEd9YhQE8/7e/kaanJtLICAyEb0XUn7btUB9du21U0xYBw/mFJaQsUgE17vnf+HswkZvpyksvpXs8t7mymtora8RI7qa7771OntfNdPTz30yHnj4HnbbT4v2baQWvlvDQoO+BObRRxdrYLqEu0849KSKHecy3h0ckAk5CKXcQMxMIRBKxAAhFKQIRqLo0zWCdnMarMha/dXuKPGAl3LuTOqo9YFEgE4i0hcqlhOlt96W29zfR96Gm7ipPbK+B5FthiFZBpFovqgjf6MGu0S7RKK5jP/QMj2Is92MPDKXBg4ei48XUDtJKZEMApb5BCIe6V/w20gUZAjqjv1OHSl3wgtwE9fpGRmgeqqDfgBH0oJ7VkVpV1JdtkGsHVbAKUTnOozmKujQxlTYjAwCSp52CSSkl4d/94L20NHOPL6Y0OX0oDHXB1+hrpKfPnUmT2CQDA/1pcWOLT1NfYBRZujxvWZ31duoFfmvYaKsU0kdKitshEH19aQ5GU8GZAcxQu9ZmcfECv03cxCOHJpG6fWl9dRHCx2ajjcGolIhIpawRUJCVlev8nGwPl0gEZRCJBl4hQSCMiIMoIUAkubnu3cMYpcMMFqqTzt1GdVCPMdIbI+roLFWRbmwJR9qp7rTR96dOnEgjcMuVheV0+dXXkCotxkd0Qu1pzd8Pznfkuc+n0QPECECq7q1t1IdFBilRLhQwQdp439gw7tdhdPQ53KsLRKFXsBFwFmhu8J3eXtzP6ucgzQbGbH8fah9Ejq8Mm6MZiKwXdPLksUC6Lbj00p3FNABS9fDewrWLaeb65UB+y5siEVFv3RZ2ScRttI94bhfEXZufI3P3Znr53/8u6t84KtaxdJhM3Y3NVjowMZK+dOZ4enZ8LK1tbqfF5Y10C+/VBirWJrtl1CDaUYj66cOH09emT6cN4Pf63cX0o8vXSGw8kU5+4XnspcU0d/UqZk9XWrl9HVgRXd9cTzPvv4rBjro7PY30egpYblBmN/YLHjjoIfoRiEi4buUx/3q8/z4cIslQDciFC7X8rX6uqlVKEPTcPrw7B597Hr34CGnbw3Bb7JStlbR4+1ZaunsnHThzhtylEWyCIToNYukfpoNyIM4AWg1J0Y8rdBFde3lmHhtiLnt45mZRtbATSNabnz6QBuHgY4eOpUFSPYZHxtLgseNp/CBlKZVQz7qIZtfgwruVrnT33fdTP2pRz0AfwbmutOH3aINy7NqFS2ljdQnkWoHYqI8b7TKuUccWaqBKjWErdCER1pfmqcOdVMH+aeCy7cHF240K00VsIkqT+NhlIao8b/35f0i3iGloOxx85jNp9PAh7Jru9JXTp9NpCPmJ8RFNhnR/djHduzub/rf/8X+I/CzTTzT4Gzgwzj79dDr5T/9bJF01DfbW0tcPkB7f6ElXFlfSezfvpa/8o3+MK/jldPXlV9IlpFskM3a20n1soAqBTBMqD3/uS8AFRtIh5rS+nJavXkhbEBcdEM39efvzcIikgGKWIoVoBsnUZ2OHSCSUHhCrdwTVYXyScdgLaZWo8vrCfGrN3Iezz8BZ59MAakyNQRN9BNkQ/rA0ERa04lRVQDWkQllDpF6YzZrTw1GvRFjyn+q4YRsQURPkbPT2pTpSq1onko3xvzZ7L0bmbRCldoyFQTs9auYytTB4O3Bw86NM09cA7652pw2QX+KoGR0nUJfb041K1ICT41HiKOKr7kmAJifWYAbmQHVjW4QnT5CIcBC9GbrbO510/a1X0/zNa2l1fgbnQRO1x2j7Tlq+cyc1jkwjmTDwd7rDm6fU0ZvVQmpZZ3ef1dW7jC2ysbmT1jdUTysJ+kjjqEutRictDA2kOaSpA7Omz51NN996GZhZT2wsgrCrqF49wGcDabpDncLty32H/u7y2y2qvu+Y2/F4E89DJZI8Hhri0AaRQDyil0sgRtH7UB0GSZUYgtNfeeWVdOPNt9K111/HZXkH71RGsoMnz6R6Tx+j7kiniG4CsTTeKc8LOeUdNebwdNgGczfvhDrT7JqKyPUQ35gkpjI8MoqqhBGKGqO0WLt3LS2uL4anSCQz8c9YgR6dXp5ZB1klELNqszsUNys2kCqINsXoiWORpRteOiRaBNYgWtPXtxnotLlG+gjvjx09EanrZtpuSii03U3EFOU6+Kg3Ibw3v/1HuLHnA1mHyPyt4IjQDXz3rbdS99NPpu5tPGhrKYhkp4r0HBvFOdAETsQvIBAnclA6ua3jFl9FJdMI3253pbEBVD/iJUPNevoX330j6n8K1/Dbf/JH4Y2ToHcgsHUYk0S+CvOQWCgYaWIQlvKVskHY8Qn+8C1+l0RTXn0cjw+HSJQahYoVxFFIkTDcQRJjDkqQqVNPgOQpvUxy3d1Ll1GxbqY1/PcDGKbDuIMHxsZAWAYFwcHb6+u8V1c5zuzMcQ0azSC8QTKHoTYxUAfGRgjCraV+YhDNZm86ceRgqkOQvY1GGj52NK3ibt2i87epQ2VkGo+YLmi8VxjyIorR/Df/+A8zYtDj5lINYdMMTE6lIYx7DWU5vRLKOIKbiAK2phppGhrrdQYyVccHuYYaRD7ZxatX0hbesl4ItU06TaAW0nAZiXn7/Nvp1vn3iV/cRoUjiRGkT913GZv+7QiI9uCF+1+vXkxPP/O59JWv/WI68uS51EE8bLBPnjxNsPEy0mc2vH/LSII3f/S99L/wjf/8t/5xOgl8T2BjtNaImkOMV1ETNfA3VlvYcSvp2Oe/nG69+QpesQ62zUKogzWkrcxC20/K79o1EOkxE0UmlEwwcW7b3Yr7+cfj9ffhEEkBo5JQwCYwCXVFCYIEqKAqDE1NRtqECXl3Ll5G918JTj15+mQ68+UvoBYZ36inbWyUdaPleHtGkDqWWUoRz8FZfhvTgFg4r0AQunGb/X1Ij5E0hku56rf57pqj7rANzFeqD+X8KvU166V9s13B8KfuB86eK4gc1zKqSoM4icE8pVZ9AHUNO6hi+jrkoSpiOse9d97Bi9SPtOrje9kbZsLgrXffSrMMpd2C+AyYGqwTuagqHqWVtATXXqZtVCKNYXvoyl6+exvnxTpJnGtpl7bcR/3bUToAq7+FbdIg9R2qTwO4kRdIz5GwdJGr0nl+6+L76Qff+dM0iz3U88u/mupIkzs4Jt67jD3FIKstnRs4ISaIpi+i4i3fp34VcrSU9uxKph3UL3RS6ERVC0kSNabS+4hBfqWEiWvcely3h0ckstdgsR45KXZjJOYbqRef//7LaYGM1ZXF+dDpBycn8MIcSc9+61dINK2kdaLSd6/e4AiHpVOH4eakoUZfRAfxR0IR6Trc14WqgW1ArbdRTwN4egYIAkKaBBwZx42urQNAgqiD6CJV5oaoQRCLzxhhP/Tss/FMxHOMvaAhqbMvgOwN3cYQL1+Ob3dQVTbwit146QdpiPSYIRIJu6g7lUmLGOOv/tEfUl9sFJB3A6Joh52jKqTRj+uW78pDajCF8SeeDIl1/tt/Qntb2V7h3VXUwXV0rVs3r6eTv/qtNDE+lBoVMndxRtSQkBE4VQIJB+qzCjN45+XvpzWI7MwLX8Uh0Z2uQ3jn33w7DU8fSZvASlXu0NNnI2Yj0wjGI1ALeCJSkaYQNIHQEPeUbflB4MW5v38etk+XSASyWxw9F5H22SNiG+rRFp348h/8AVyUnCqe0Tb5/N/7+2n40GFeAWNQTc5MDqeTTxxLtS9+Jr156Vp698oNOC7Rbwxm86PMa+qVqyOVRGZjGdt4kxYuvJVOfvbZNAICTeCh6odYUNLg9q1068VX0yHcr30QSZeCwK+XHc83NzDWdQWbJ2b5pm/oUDAvTBXrwLmn023sJqgpvHBNrt1+78301p/8O1C5mu41r4XDQPXFtI8Vxo0vXD0f6qWEK3EiMvxwbBVUs8A3xNeh57+anvq1v5smTp0IYr/84p+lNTj8poFRuHWoaLi4f/d/+p/DFnL6oBuvvRJeNmMrEmUuDMQF/ms4QG5d+iD98e//6zR75VJIrU3GvS9DJN1KU2JHa9MSWTPswnXqG3Wkj3q4v7O5HKri9hppNEoS4cSx/EZIEGr187B9ukSyH2JypILbBtFIOIpzuKyGez+SpKrXieDh8S9+iXEORpGH0wFUm6cmxtIIHLIPAtjcraQBcrD6eff9t19PB089mcZIwGtAIBXsC0lRjrcyAzKD1Nolg5NjICbzTkEYy3D5XVPrQZoZ1JBd7JQBiNHUk26QVEKx8x0Suwq3XcD1rKQLZKD0OgTTRwQ9PE60oQFSmeukBFvGRb0Ad58nWbAxciC8TN1VAnGUpedsY5G09l0TAykOT5hGvxIo/kfFRTLUG8oaPXYCtQ77i5vrxEv0VIUahWTL3JsyuLd482q4tlUR17VFwrDOTg45vlCX0UjYS7TnwvdeJNsYFzd16sJLt7GykOq7zJ7S1Y+BPkNbyc/CEWEyqB46E0OtYzA3ygki/giBxMWSQLyXH6Luj+f26RMJHS4C8WfvCMyL3xIJ1+kYDenmFITAOIqjzz8Pd+6kQRIET02MptNMWNBNsGsLA3JWl+bKCgOCFtP9G1eZCAGvDnaCU+fsgPyqR6aQrIEwxi+2UQ9047aJRq92FslTWk6beKpa3J+/fS3tQCRbGNO9IHuFOEKoaxqmEIvjVGL8hJKJIKdE2MO3jIGo1nSLhEgxvWsSli7TFRBtJbxCGOq5oaAPgUc8VVut1dxe2h/c2PeReOJXSAZxKmAF05g6EFzdS5skE5axD9u3h6jcMwcs+Dd/tG9yYXB4ynYTYTXOde1u0a77F84TS8K7ZcYBRNDGfor4EkzEXDelci+B1pDGtFeiyZ43JZ4dx2YFfo63T59IBKaEUCCMhnFpEHoe0gSpoEF84oWvp76pKaLPO+nJI4fSSeyHE3DutbXtNDM3D4HMp9/+7f89zYGMCyBHE0P8NQJuuxiepnksg6CO0245DkK0K7478RffCQPejhfZNORl5/N4ifpJWBw/fhtpcwgEyY6EblRAbZJ6c4jg5RD1F4szYgQyc9rGRbwVUWgJJN+/9e4baZYyJdI2BBYIT7uNLzjJAg+S0u4MJZ4ax0B1stjYKTnsM1Po6+RyTQITXMw8J5E4+jImcpBIhClb5thZ5Qnpwr0oLD6QnxKto3gIxBlPRPSdNmrdNvEb7myS5OnAKnPiWsC3jo3miMhtpGM38ahu6mTyY29/TmWxDR+7xUc+9s5jd/HhEAlgQrKDBMLLk7xn49CgXDU1xibSARCjn70NBzPH6D4GZR9qyQpE8eqPfpRef+WldBVE3ADhzM+qokfXG3QoiLSI+zSi3nDMUDFEDRGDey3UnDyeBKTTx99W1VEVQTqhY+tVamMIKykkEFOorGvg1B5K2ohMeLbCdmRchEvzDb+1xEClVnBjcqGI64S9QUESiQU6IZxz9g6hHmpXqEYZTQ+jvURsPio89J5VGTJr0E5JUN7WuSDaxwEJJZU9eL9A4HyTGkY141iqQKpyO0yGoeNgG7VUx8FGqyftol45hKDOfeFg3Q1gqgorvSwrNALL/DnfHgqRiGx2rYiVAZ2P2YjnnA4aQoIM4DLto4OWed6EwVm4dYfI9+Lly+mN119NL3//r4KrgRaBSO0WiYh6j4aq6c67b+OhwTUaSAUCB4HAWUHQLdSuXbkyHiyj1Dtwx6gRhKKev42+7xh5RABJkdmjhSspyhIB9ZLtIRnlugXnBhmlD6WSz5kHphSpdNeiXNsncViXeJ/2i79KCidlMDjZWl7Iz0ap+Y/woqKoiOsRD8o4Hxf3ED7iFEEs2hYZuvkmr+ZHo4nlaS6ZvzyrdJJwTaXvws3dBcy7kORN6hlSHsKQQ6h6aZtYn6iTheXK7RW39zvuPbj8OJ89FCIRYHtcSICrZgXkQRi4psNPp5jUocOgoYXb99M9kHro6NE0x9Sbr7z7QVp476108/13SMZjzDodG+WBYFsEFCefZLI2VIQbROhNh09GlkEwkZULIDHPgWwZMfiejoLoaNyjcHvHcdSIV3QzjXqH/Ks2Q3t9vvRmtbFXNlDftnGRqqq14baW7cztMU2PKo5pKiCWE7z5PSXTTodEQIhRx0R8GwkXaR1UfvbW9ZjDd2B0QnIPwpHIQiJYZ2yLdXLNVu/NxBRAqj58IJiJxBPTjUqd7OLmhzYvFDTz4ze9J2Gb1Vyk+MtUgH+VsmKMiCqnxM0mwThORomipyura97ITK78jL/3to/e27vx+Jw8JCLJRCFyZhWrOIIg3XCwHjxTU+RaLV+5leZwc1649H6a/tznQNRVBia9meYuvhcuzDBe5epwexPUdcnKpdXdm7y/IVcWcUA4kcH5qWq4fCePnARBRT7GVqDf26eBuCDjFtJngdF5r/zJ74cdYqpMqGZG4DFmDUY2hyeDmMPAxWCvEPH3vMcd41duq/5+8wIETSykg4G+QsRcYnVe3diUNkiqnQ413yKISCX0sJW5UiK/9TZ3S2RcZ5jx6LHjydkVV+dQ4whESqA7lNNFm4VjSLN9uCdyh8QS3HsYnE/2hI1t51K8K6yEUXOAIQkTJDM+ASwbwXxsu+kpAIoMgwOIduynLRwtwCOYjECkDnmTaPj2nijjagB5rxL7avnon37qRLInQYSNgGPfu+ZP9PQqLt1egnT30NGXb95Ic3hgttCVTcs2jVvXqoarg5kkAtUttyAa9WU6yzhBEGDGgOhMXbcVpuMxUBZeIYhEHT8QpODcVbxiTTxqo0guO1qu2QMh6L0qiUTEiXOlHkQnYSo5GjwjMioNjbgvkqOlCtNivLq2jhIhe5xARtrqtgvidVoaz9wLt25OKBQmSpJwzdJG583qJ17keJMD555KN15+FdtqEacGkW/ytoSBLtyMmAVo4wNeDyrwa+x+GKYS5fNb+PNPx4Twcpcp9BDjceCXz9ueAbxrvi4z6UFN3VhngBXlKvVVTa1v9CPH+EQci0ZSSjTYejyG26dOJGU3fRKsVEnoFTgVXB4kWEPN2IIL38H/vwlyb8I9He/h6D87LDxJdIimQMQNwl4Aqei8KIueVdcPI5nOdcK1uTu3YuCTiFFBUmQumst0ft0JOOhXf/MfMWTWCdogEhCUwuI5MSXjXO5wkXMdD1oHJHddjxaEsaMdQj36SbrUibBO4NEcLY34Fq7qNoSsl0iVxVGSfCjK7+rCA6a6I5ZBdNG+sG+2I02nSdZyLwPAJsnPclzHKrAxVZ18mXCEWKPIgC6AG3haEggE5/bAdoF4fIDNlJxA8lCt+DpE4VieUBeBnaMpx0mJ2SDrehfGYmJm2GyUbRLlJiMiQyWjvGBMlOdk2lGpqATfsR6ee3zMtk+dSD4EnxJeBSDtKNUZ/qYF0uIXiV2sMI+UXGtLjgvy6YUB72LLUV1OLYeL2gm6UUVY09YtNu5Fx/DDC3Yg9kZOMZH76bnhFogxcvgJhumeZrjuULr09nvxchP7ZhQuqqRQBSqrzBu8mz+gbeF0RUt43VpkCSgJXQtEtatBCv/4iZNIQcaCG1tBbVnHltpC5PjPciMwSBlm6rqpPDoM2bZywndIG3n15UiAbBJUHcSpceIrXwF5a+kD4BPOBZ+lfiMEHbvwAMJD0uL1SxQGsGi/MsZjqF+cW/ecogOxhq3kE0hOxrs0ULWazPHVPzYIYZP6v1EJ9cscN0dtGpStAUe9XsZQ1sh0UF0sJVFkd3M/6o5r8AHMonmP3Z+HSyQfBVfgMcAGrMtw5FUCfesECjfJTRJhOvR8jN/IXV4gqU+L/OZ8mSCYiSUmUKDjspQAOVBLVM00rlWPMucUAUlxJ3Do+JKBCTKCHd9BpqvSqofMXxErBl3R6eAIm12e34vqcy3sHq63iN7rQuVJvqNKl9Ur86Vcp9A3JWzboq0jcuqxklAyweFeDRdu/kpWAzOKzTDo6eCzz4VdUutj1OQTpylniyyBC2np9lWkF+PkQdpDTMi3zgCqJYx800tkHkEoHJUimbFkdcv6hxSxbXquIOLxM+fSFJJq6smz0R7LWWWwmlJdAlGyKrFX6Z8+CKQXV71erxjeIFyDOGRQShM7lI94fIy3R4tICniqi6+SbNhCj9+AaznfrddEoEg6FFsBvLCXQOSIci2JRCRwcFV0XMZqHvIaRmeXqe7FcNNSteAZPVp1VJk+0i8kkBy/IMkR+yRzeqYcKokk+jt3eiB2rkrgYUzepueJWkmIYYOAvDoUHHClF0r3cQ17KoiUepkekpGM92hDJCOC0DmSblmZE89fvxKGvXNg1QjkjWAzSWwjeAFby6hdmxUCfP0xqdwcw3YdheksMxKJdUHvs1oMcc51E2RRB5EXGGgLdpNhPHH6TGQ5O/Lx9ntvp8UbN9MyMzXqXRsgLagCYVveKvai+XA9JnMGbARE2S/UuShfQsn95Icez+3RIhI6cYOM3jbju7fItZojzXsZDtZGzSrVBNEmD6pSSvBD+wVOpgE9CsIc/+qX05HnP58Ofu7ZGPIawTkydNWl8wt53iiNU9POB6aPMCXQGPZEH7EKDHRUJKcVwtNJThPSpxtp4FQ5csii46PX7W8RgGCj2Ke3yZGDIwcOBbJvkf6yxqQKDpjKBA4nB8Hk1jXHmsCRu/GEmVrvEGIlUHjaMPid0MHovvaFszJuQdiOQLz0Vy/i4VtPn/+H/zWOCTyAZ86mb/yzf5Ze+zf/Orh+g5ScwenDDC2eSgeeeSrdZKZFPYKm7UhqDjOAUtMHf/Zn2BFzQWSQD1F1UviJL42dPpu++U//CXUhbQeHww9/51+F+uqkfutzEArl9mOzjR44HBNRqGY2B8nrAo5V4Cthq86WRLMr0SBXEWGAap+qGh3HrcdkezSIRKD5P4AnQIFeaWgGJ/JCPPIAUQNjxdOs+ihFNDRj2h/UjCZczyi1NoAR80BoOlBXbUzU1kB6YG84z23EAyQ2vrm1tgT37eDBgQvryqWjXZ1BtSQqxrnfVIJFPUEAhxKvkLC4RhqM6fodEHuDBTsXcRDocZNIbI5H3w2jne/tVvXO4XGDWHYhoCYE4kwsDtYSFptIGaPgxjGc+HuZcfoVxp/MXPxCjPl3IJhDkE98+QWkDMOJNxi41YfqyAhDZ2zsHxuJPDVtHjOPNbJX8Awa5Ye62Y2FMCxh+igT0Z0O5tLgfUdhrswtR3JpHeajpF66PZRGxsaQrhmmzgVgW2ZgDIOHDvA9njMQCZxcgDQTC3CKdIUAmgCIfnzc/nzqRCK+xlYQhIATIdSXY1dtQl/HAAmAC+w93bYkDDshtsBe+hvkBymGDx2JOa3k1toWujHNy+rwvIQgotcKCeKkC41hJAhEkqcfNV8JdQeOvok9ITLwJxABbIpvZKKwBXa6FUCNg4MugCgmMsZYEIxn7ZEV8pvuk7flOz2ocL3M6aW3SmZAw0LFCu8VS7NVMB2cFYbsSmISLqKDpw2u3NLYj7gKXiRUwnUnriDAeevN10MtHMCAl0hOvvDlIKDZSzfDs1brIVMBAhk9fhjYaR9kNc5kTjOUAxZ4rDzqkRtHIh18mgkicAb0NJkID8mxeGcmpm8aYVHSXpI+m0ibcWaq0VW9wBSpQyz7sEpcZ5Y40AATWzgbZnjFgLGEopq67TnfDtLgGOfR78Lu8dk+dSIJgshgC+IQUcKglNvyrwcu2DvY50BsOKITP5MiIaDFSwENx43OKIEPtkogdYzIw5/7AhIkr0++OjMXCGdnOYGac1pJkKpuQ4ePBVI6gtDyjEH0EleZYBy8ksfxLPcuXU/D5I0poaSIUt2iElF7DqgXHSZpuxtJgX1jo0wi/WssQ8BMkLzfTwbxxst/HS5ekwXHjz5RtFcFR2kpM1C/j4ErHHVKkETZXyNC78RwDrndiMCoGccOG5bgQff01u//G4hlI+IlJ1/4YuqDSSxCpBf+4s/SnffeDQJ0LPoIC++YodwLcitVTYxcRYWtYWv1DuXU+4knzqVTX/8GM68cSJPHDzLr5T1skfPp0vdeBi7AGvXMycXHDqCOMSTA9l744YvpxOe/FvEjl5Gbh6B6+2FMTlKHu1sY5S06bd/v4vJjdvjUiUT4BDMpJIc/9ghFzw9czITGXtywLnnmGIZAUDi0mcFyvoPnnmV9claIxXg010mvjrMvDoPkznVlqsoKi9U4q0l8S/UFjmpAsIkny1ypbq4F8tOPIp9ItEWOlcb1NuM6anxHZI+UGf1V9rt/YrcVEgkcH0nSwxxfFcpeYkBWDY5vvphjQ/y2XHkLyZIO57hO3IMgQgXzgaJImYDnXley9TOtUUTkQVKJyhWnzERw3Aq1SfdQuzYYsWke2iZu4MUbNxg8dQX7w9wvvXekzS/dD6JXskaSIp+T4AawmyZOnYppkSafeAJb5SgI7xCBdrrw4l+zfPVF1DdUO9zoO1tkU5PBUOHdLZ7RKTJ26HhITdNjeprDuLsnqbvSj3YiPfz4nvTn3H9xTaA9httDIZL9cAo8oQMCo0CQDkgl964QiKuBFFUMajku9MN1DV+mDwKBKnqhiJ7X4W5uTa5JUCK5HqMgkmImE5FZD5LE0Qu3i5wjsYhNpPR7XUgQ+h8Eh0i0C9DV26x9DgXzFN0sAsdRJPA91US+BSH24t617JZcVInFAzH1kHEbCLbj6EEkmCkk2he+G96rUL9sfBQZx5C0SAaJtmu3FwMapsFuxD4W5uEp1cjVu7fI41qOaP7KHSaGcLwMkf0OKlrkhPkd8sVoHYVndbOJeimcpp58Mly8A0yHOsysKEruDl6z+Rt30u13z4cTAnZF3duMb3dCP+YQpq1LToiBWmVQU4mst9E0Fm2s7LoWTg/2AFq0al8Dy9+P0fHhEYnUETtqFh0Q+jnI3MEbJPeuIcarTABdXWYMOsajC8U4n+0mSHDlrZfIYXoCdeMcKSQTEATDWFElJAw5u7OKLFy7CkfFaMa2cRKJwXHGP9iZgZP+EclFBDxY2hHUpcM0n9oCvhdLMfOb3s/dKTOUOtxABEuQoByvoorVQ93k2L0QqWVqxLfk9JZB2S1UqJBK4Kzf9Vq0GWSGsuPcypXEF8RCUDUTbQ4UmovmeA+nXXKycOs4f+UiOWpkG6P6eT0MaAhe6Tt5gumKkGjbBBc12k985avpMEOXD5wlYKrqZjuoxxoB0JmL19LVH73OhHkMuoL5DB0aAqbL6d77uIFv3UBtZOgAZfYPj6djz305pLSxJB0Nd69ciqBiL1L8xzdhVsLtx+8+Dlc+fSIpkMNODETgd7ZJCgSRk8PB1khFGWCchdPs1AD+7KULobro6p06+zQTOH8BY/NZosOj2B5wTJB7kFGLd977IF0nr+naj16KtHJdvb0ECp2L12UbfC7UN1dhgqCW797FPsEuoD6DU0dQ4+DGqnxw7h2yiiVgu1iiUJrY4UFnXrOuGLKRj8VNo8+OG1nGq3X95R8Gkegl6wOxtJtUR0wStATVni7UmQ0msygTMEsC4aNBRNtkBlgvU2ycnUS7xaj8KrAxs0D7Sem2TcqLEguBwRAAuL5Egmdunra55EOlzgwt1EOiWiU6fhWXsNJBtcz6LN26jRRitkayro+98DV4CQodhL9KcqlrMfrNGbKuE/BzwotlPHkSekwt5HcLT5lzH1v1n7ft0yUSIQgyBa5JHIF9mTgkFCFcShQXsRlQhQK5R3dOwcng+CBgH+PbDz3zWRIQWYKB8e99zKDiBHEdh+PCve9fuJDmrlyNwJtzA7suYBf2hfaMWGT5VsC1SIKjgwByScd+jEwfFfP5DukvSIOchm+dpYjc9dbZqrpJQD6vOqjXzGi0tsPCtWvEaG4H8puaEhMn6MqNXanA7CfCAQIIVRJpZzn8iXsfKp8rEV/hG9s8I6FIIFuruI79psTGMxRlhYLwnfurG4lmen8FO6PWpI5I0RWMduuvCzg8iBI56qIzyhuEZYVExCOGOkxDeJmS00dcZJM0lRm/rZRlDy8eMDL9J5hOpP8z0TjEFa0o4BONKQGXfzyWfz9dIhFEdiR7JpCsZsVyxgA/qx8gKcBvMTzXoFitMcb8T0/SaU1+jzJh2ueYYG4CHXwJDngf26EJx17Gj383vf3vfi8tXL8Ogczi29fghfnB3fVeyfWwBLgikYJEYJVT/5gSvgZib+AIOIqEiiXNIJxsRxQTKFgQ8ZByK5E4kBqkazKsuA8PknbP9Zd+mGaQegbr+siDqmMndTOi0NkX98bNYCfEVKhMp1pvMG3oKupSGPvCIMOnhIXwclcSqJb5nCqgTEQvX514j+1zeIBwazO59c4OTIPKKgVyNL+WVmnr0u1rMdZmiFy0I2cYd4MTYxfpUkHFzdOhbqbrqIiOlBzCuG8asJw+zGwyu+ny92x9lmSqtGvzM8E8lGb46YAlQwhwEy8tCqfchqi7r/H7cd4+fSIpoZVhGQAs1S47P/YCMRZIix+Eox/5ImrXwFn05ZX09h/+BRzxTuQS6SodOnwIBF9lIocFCOQy3AwSQOK4LoncswN3DxUHvdx8okBUvr3JeIhsfxCPIDNXqXD19dcitcPnt9ZXQs140NFUFYKQa6vi5I2CoPbV+0z4wD537TJ1uIptsxrjSpwl0hlX5Li+Yyp7JhSaid2zGVIsOwcycdB+OTbtd/c82ykGIfVMOdwXbq0aCKxMs99sM1E1NpzqkQThHFmFUOGa9WSXAOUZHWJG2Gcr1OeN88wMoyRhYJoSwlbZJRQCUzoTrtyxg8fCLukgqS0pp9kQmKTuEoeOhBqeONeC32rhRFidp5goxZI+vHn5E259+MFH79fDI5L9sAikEIhAMc4zkkR2LZ3q3FbDR46jGrhOBsl7t++Eoa46JFfTaN8k1yuMbRAmOC7EhQ4UeojGbzc2iL20y3BdpYXXVBX0yoiMcuHl+8wxvOk6IiAiCCgS+c6Pb6IM1z2Avc7Uor0h8VoHDdxeDGUJQmSPgVFhMPgCXBfksm2WUR7jM/vanmEhwUg4pTeMdziPgU7ICuunM8NFjDJRWdcC2Tlzoj6va6DLfHa6YQwcI9UFj5uw09sW93nc2lFp1E9mlMEuCxuqgFFMzicx8T3h04+qG7YdEtb2uu6KqmBsfHNvK08LWAW1e3P/M3sPP5onD5VIogNLINq5AC5fAznoEN25HWyR+ctX0jSq0PYonRpEYa5TnindAVga2XqmqhiOrj8iwrtme3BX+4NydGmG7x6VK8rmWyKYmzgkMi7ddhQh5Uhc5RaYU/7wuVzhErHErFhRK4h5FmmBaoRLuEFEX4+TEW7/iVDaQdKwdpK5URrbodNzPzbetR7xDc9BuiBUEFuDOV8nVY334zlfsm3WKVSxbIiXiGgdo54cI71GJK4ylKD4Hne5/pEGcs866aELSWs9+GdWg2s5+i3rPw7T0qBvAfcKdkuP0y9Rz70t6rT3K072uvrDlx/5Xw+HSALwIGcA0g5mF8DF7+Ce0ZFcpxMV7T3ovAdOHWfo7H+R3mEN8ZuvvZ7unz8PF9f+oLNBQpHKeaM0ujWI+8dY59D0EspVAuUFMhmgxbcCyTja+UNMH2ROl9OOBoHxfEkED3Aod7FIks8ycvmcU466ei+YHPlgroniMOBAXh7zDW0G62FU3dT5QEKI2bt7WxAt3y6uVkm4zJ4jXMy2kTYJD3eJJyQJ3+xgh8Q4GlRIs1iUYNYrRhsqwXwF+Kh++q7t97LnPBXHvTrwrJLGpeUaqFF14lT92CXTzzyfZi6zDgmS8v6Vi2RCsLS28wFgE/UN4tWCeSnZaaIF5zI5RF35GFYg38+tzUS9v92fcL5XqYd78nCI5CNtFmjCMzahXMDMtUJc/m308GE8V2St3mcQFqrXBp6oDsSwA1LEeuyBCEwHhBSQe8uulSJ6YTx32O4D1JMAAl8DUUQoJ3pwE+nUt61PICTXfLbcSvIIJCgryU2j6vEtzveepwy9PW75Pf5G2/QO5aWy+WDcj8Z7r9jKU71v8Y4ZtFzchsBE/ngyyqLM8ErRTLi7qeuOiZf5xNqRYmWJsBx9JeonA2GicLMXzP0KD5kBVeyZOrNEmtOmQ2No+mgwKIcVrxGLWSFVxtiIozvN1jYI2zCFKDbhSv2sg8c45/uquaYTIEWteICzqHvxItcL2OxdKE587hHYHiqRlJ2dASes9gGFU+MO/WNjpEGQmg2RrDgQi2zXTWyAiCxDBEoRuaToFqP/KEwEV0KESgb3rOnHF0kEuOoM57yWj3SqA5+MLsPmQ50LbPJZNp/78S2/73XHt/gdZ320xHjBsrlepV5+M1y3/HaQsdLM4buU4OvF36hZ/C5hICi0F/gbT+XzXG58J960jKIc3LqqQdtWg/ci9pLFBRfcbL/PgqwY9H1k8Q7i5RplSTlTcCrEYkx5HyTlxOlUXdzUyb+ZSQl1ssUiRosxt3AL54kBS9dXcbxNtUdCzMRuK4I4AFp8S+Dt3/02/3NcKNfb9sZ7vBttLwEuAB6R7eEQyScBQICxi/h9dNRzBAsncANXyanaHh1Pc/TFTYhkBpdtL/GBqROnSa4jPYPM2Ba7Q3+NS6g6VfC62EF6gNpIiJwezm90ePshOoRvOW2RurWJg6Z/yL0jLYaHoht9OLay0+h+7xXXtW86vKeh76V8Ty+aq05BeD4LwYo2ZbPjOdGoLKO88ZHvlJc9ikolYcWv/TdtDxJlp83EDKh6ticGmVHe3jesh5yf4OthRjj+8n/3T2J0YwdGhHBAgjoATMKMj1FXSJprqnKrLJr0yr/8nYj9UGKoWdNPnsIOIU4TY0VoHe0NzxtMqYvM5pBmEKvXwwbkI6V9lWGfCSugq/pn+2iT98q+CYDtb2fuiE/978MhkrKZAKAEzh5ABEqx1+GEhsoqcn86uRtEXEeKzOOFakXUGM+W7smIdxAsDPVFTBU7RBA/RGfQCdkFpNpiZ3m9QFq/FeoadVFV4NlsMIM5FsDt2Hyu+KH2kMsGkaiThq4dG7lVlCVRxAQTnIvcUcTeHx0GRaEc4hPFz+JDhYMh//Jv3I4PgujUMddXHV9izMOWIx7CPWM/ttE6aeArVb0XqwzD+V1n8elf+EbqYUb9DnaSa9x3tsx3wyVNZvGNl1m0B4TWQXLiha8wS/4b6c7bb0EgN6KvBkZHSDA9l85+4yu800qLRP9dGi+M9qIuwiLCSgKacqKRwljJaNshRKXJHjF4z2e9xzFGT3Ie9+27cvP+Q9geLpF8XIMLQAigCh2IpoxamwHmMUbJkVPk2HeXUZYGaiQDOu1PrM1eANfo9gMktwMKThadkfuj/HzJ/cP4BxkjcU+VKFNC8ZgdlDspEFv64VsiI92NCgPCqtfjYAhiATGD6HhLooiy4nXaA064hceLcrIOIgF4MVNgjul4r/itCilR8FCe6yojv0ONjZ+Y9hISMDxzua45j0sCcUb4JomRLFzENEFHPvscy96R6InUta0SyDoTj2v33X7rDb7Dt/jO1JlTLKL6DnOdvRZJjw3S7k2uHGGmlmlWydpgPrE20tI0onAbs+qVDKIbN7s2yM6Oai5bwBKG4aQQwkLEj2c8sgs/+lpnSriz+R1LansPeASxlH3B+5/29ggRSSaEEvEkhvCWGCmnll53NxcquKdIE0MxRBwdS6StSBixZUDGO9EBuEadzZEtcDLOomu8G6qYapoLiroevN4nc7h8uiQUETSwmI8FNy9LosBuhtpW+/PsIaazy8GhmkDaCPAFJyc9XzWQ7zieQ6+bu9m+/g5Ejuv1QHrT5WNded7RKNfIzmkoLIpKFN9zpYSqpU6KICAJU2BISEFM+aga5JzGtj3fD94S6pkq1uyFq+nm66+kd/7DH4bqegoJcpJkyJf/j3/BLC0/ZMWxyxB2FwR2kPy4Kd4j5YVAbi8zwZz9xi8zBuVKunfhAyYHv0TBtN06lPAC0UM6A9c9l7aI7nX7VEKQiOKc30qe4r4DzpDPxT2JKRqQ7+dfn8rfh0skAgMw+E/AdMHVFMUC0OjuKhxqXk7FtV06aYX0i9rBw2kKW8WAouuau3yBrwDt3DEidslxPcZudqwcF0JAZ6+by8W5i/H0kG3svbBjVE9AYjtZ2yUi3AT+wnPEs6G+RKYt7xaqjDlOgYS8VyGVPLJr5cTUN6tD1qFAYtFUBIpjdsf6bhA9NOjYelNYSpWqCvL7fCB2UWYuS4RX7FjUvjLzFS8WZ+WxwKvip2APY5x5BG69/k668J0/J3/rBkHb++m53/gHIVmu/OB1kkXfijUqu7De67h8G7h8HRag1L7x9vuk/PTF9EYNhitMnHgihjBcZ43FLlzBXRVnhNHe0wYkq5sUGKWNjM++lTAkktzXagT0n7t15J7HXeYTy94xCYT2er/cAnfKHz/b48Mlkr22ieVsNLzkLurE8+i6u3DaOsNUu+mYbjj9AKMVD9NpuoaXGEexseL64fl1ETfEPRy1RGi5rLMp1lgZN65BEBKJhGF2raMf97iyxKKkCsTMkkHVKTsCfB+uzmCtbo6W5W+JTQkh8vo7jPQgUuokN7dqJaJzalXDbVvgRFl5cb0mt1dq+g5vFmQQvyK2EzgCjGQaIEw+ZmRTbcqxH+/JtVUFeaGAaaib+eNxb3V2FkZzP1SpO++8CdHkbF95jcmeq7jbnU/LDILSoI6kT7x4Bmo3CCI2tshXk3jxHmrHqN6ZR9bxm9RPuJgzWQXOzgBpmnnN/wAAQABJREFUeo45cU69pHNBAjIFiLUxOEea8M6evQZAxAUZZ0gT9bDimkfb9WltXeee/vyn97X9rbKh7AFIkEpO0QXChV4vYkMUTdy/I8x9+/Rv/HrqP3Q0AlsxWg7gzjFP8Pz127gmlwJh5egN1iF3hJ3p9KoxDv1VGjhju6vU6kNp44IF7oGJwtlqxCbcub/3u7hcHj4JSHRl8X7B8ctESMtzEoSiM+NQPBrIzXnZzwUoNGukq9jiURHEnQcjBQfVUxtIG8DhvKaVxHWPSlw4tAgcEz+AfM4OGc+LnGRJR4SfwVK6u2+/9VYMxXU4rpw9g4HJKMYPxm/LWsFzuGtuFyn7ZggPMGzBrODxo8cZvDYRap82il7BgDHSlrBtLJZUxp60n5yg49TXv0mfOvz3frr7/ruUmwnGMStrN68HUalq7VJX6xweMc49chMpIuF7zPD4kFQpO+lndHzoRBL6tJgh15UTFzp2NwCv0gFO6zl68mQ6+5/9KvlbR2O0oj1qp0eKvMa7yA2hRQAQrhaL7qjGwJn1MulSlktDH84vYSYHgI//AF24B0p+iEBEzIhN+BydlL1iPOx1O7Hg5Nm7ldNlvK4E9D0Rz8TAuMYH/e075mq54I+eIefQEpEdZhxIzlEkl2GExKJNztllTGKD/DRjLFmigDh+n3qVCOU34zdH68iT0cgsbfhNHfbOec8UeQlBggv7peAOG9SNh6OunbUF7LQ8O4pDdSswL+0nZ0pxlGOoo93O2MLMNDA1CWV9mSHMTCZRY0nwWPN+cDTWujzw9FOp0mTtGCSO0sO5vHTdbxDZ394g6ZJl9tZu3ULSZKkiwZj9INFYdwqjXrYh77Y/tqLv8o+fzd+Hr27ZSLmvu+c2XsnCUS4pEOevYBi+w6AfPCAjTH2jjaDt4KQJNTEfIoliZIeZJWZoiURwTpdG2CRNxOURtiAuJ7vLnSUnlgMXHBqkKV26cc3fIhfliDiBoCK8cYjgdlldiLiEnVpctzISrCMTRUSfNcUkvF7UzFlLTM1XRy/HyfiMEkDC0tgX6bRpJA6N5Dx/lzASTA8QxPPYC6oPFaWAQ9g2godN+LjBTiiXnfrJWOKacwAAb+vgmHmRkQ4AQUmtgXGxkAuuZfPZkGhIM93wfiecEgzndd3HTca3OLWsQ4hrSPBaL1Ot0l9bLTKGkeYja8dSP4Z+lQFwlu9YF/sqsiG2B/i2MSukFjPCKCWtQfaCWXEZgEe9XrzEHnl4XvN32Tie+FlsD5dIaFw0nqMA2VXvtOPoLDmySOJ4dAnlFrlayAVmOj+InWIqRCaMCH4JbXYliKfCDTwIAmkTLTYBcZUhqlsgW4uynEBCLq4e7sRtOa1kM8aMm+MVIxPh9HJ4EccOFIkU9RKNakvmtnBF6qnKZTsi25ejto7Da+dvXo1nlUKjR05HGr73lpk6VERTSrjZTutd4r6RbMeo2I5N6t9CkpQR94wQIku8yp8ChuVPa7N3b+9inISt5bcgCtvVxe6GEzgkmVIsr89OW62MjIH2bXcz2I2JNuybPISa5bF5zwi/0mR7k3R/Ac5VGVIgOWVru4nMzj82cuxoak5Ocw1prqoIM/AbPf2MumTpCec4VqK3tTHLTpRwuRab+AH8w1Djmm2JO+X9/NTP5O/DI5KycR4FAB0Q6Qp0ihO1qX7tdOEFKZ5buslEb9xbAuB9Bw8HUtohyzOsiYiBv8aM56ouWy25MaoM+yazF9ohoVMLcPbgkHRgcF9VFACfo/Gqa92UNReIYScYF3AzX8klGhzuGsiq6I/NbhI5gkzoNDqPXy4GpK6tNOLFQJoFCEbPmomOcsyM9NHNURd7PNdpF6OYwB55aEEIvC8ICpTIXy0/G78e/MmlISvyCS/lunlUqpQ/JZDQ+W0/tkFE5yUKvkIvcJTghQtuZghK6bNBprWGuUSR4ZLjSS6RvY60zOlB2HwQ9fA4I0qZ+8tpjpxx3xk3VxlENwbTqTPJXq3BupPHjpAweSld+s53IaBTpCAx283RMwx5QMo6vxhMJBpubcAN5yDoIgMgREy4hqmteGPzCxzx9GexPTwi2dcauXCIVoARurQGr1yOjhXtdrszN3O0oh28DvJHj/PeBkSxOreAQYi3hjHdqmix4CZI7XSd6r95vcIChQoM2hPfokDEF4yS5xnTY11CCHVrXXUEZBBR7ZBA2KjRR/qF+opbbBIKL6AgFFWkTaolpsc4E6RPbKtj799oR35P7s0ZRfi9QGru+FquffFSie3+LM7jgAhRFQ01ipeCAfB903Xk+qps26pQcuRoE0QCQUsoUVR4qnJDeC1UIcsXVk7Onc8zd9dREnYgwUyZj7ElkyudPK+FJzENMGfA2ET2NmJnuhy2KlYvEjJaw2f0EDZxBjiRhxI68vROnkrz9J1lmU4UEpw3dPBktdKaWn/gRN0CNjZeoP2MtodPJDau6AhdjXkDWYLL6ZFSLaDHIA6lw/bcdhiy4XbVOAd4DlM1hXtZTw3PCbg8GwjAhetkrp0Rs+A9xXdAQL9NB4mKqlU+G4DnuEXekkbtg807bBwe9Itl5Isl4fkrn0vntIE67mD7UDjfAgE+pA9JFLndcdyDgaXEp/jLfR8pPl98MFeCj8v/vSbSOimdEtHf2kNBINhvLrm9AyfvoGI6TiQi2hBKxCw45naW7bNM/kFcflTktO47lC/SGmVvQBw12uWEFrqcVdViCDDZ2S1nvsTtPn0KiY/tY3B9mUklnBbJ52QISllVz2Fmf7z60ivhZneiu3ESLjfINt4imNtGzdSxQwU4KvUoSPhQh2gvbSxh5++4R41/2tujQSS2ikYGgtMRsbqsl/hnw/N0snsYEkgQeEbIXb26f8hAl3PR7kbsZI1sYQkqNoGqRJK1ex7YxlGgsoXOG6eoFzzn8sz7t71OiIu+/+N98XHPxHNwbP/ZDNUsvlB41rhGXQL5OIaBzQvBJKiXhBuBSI3rop6Za1KQ/+MZv2D9Lcp6WR5p+yCWs8JU4ebCs22OGxKEpVTjed7g0ZJV+C0QX0bgtyAm15AM1ZJ3Q9UK2PAK3zRO4rf8twaxGz8yLcYpWHdVRZEmbmsL91F/76JOvYNbn2ArhHSPsSj8SIef+3zMPEmj+e3S20fTM792INQ437311ptWLzXwoG3htAgmVs2Szm/LCLiNTisgJHYlfdk2ru8/97mfwlaZmDz43/8UyvnJihARouOLHqG0HHUur6uyZG+MQAr1SNXA8yAqkIoOa5Ip3M80p/2keDvORCQRoHpQGqR+D7CwzihTnPYQR1G3NgjmkghywzzFT+acFMq9gmtFy+yEB3V70BHltaKTaEPuo/xbZMoXcjuKp6JEi5MQskQ0CFmmsYis+XfUwU9YrlJI5IbaRQzVH4fXxlRDIZ2ipXEv0Bju7pBaObaqS2Be/jJ/eSIqmqVg/k6GawzMyp8E5jChaGL8iXp4wX/5G0gXYGwsJWZQgUhizUjVV9qjty+QnPL0cOkhM8uhQaylGwveNmlnqgnYH6bhOBR7j5BpX0w4rnoqNwgYFFC0/vsIoqgh7/70t0eDSPa1y8YKhugckVVssmNUrUAqoBtIFETjNTpDANcx/A6w+MzY9BHSI4YjsmsSJC9EYt/EiVN4mI6lA2fO0TkiGhwJBAoCsfPQjzMhSnzGWii7xA1qZGdTEXa38pyj/61f7PlevutfNt/zniaxMRwJQ4Jw5zsR4ffc9JfgzF73WTmtyCwuCBEQknrHOeVtk3S1A2d3zi0RVaAFw1AS8rxtdN3IYBQ6Q9Dz96od5UXBUcXsyjWDgKwCCCMepNpKknxetKFoiy9p+EsEEsi2i4yGrdNh2PIIfcJ71H+bIKPvR/8wPZFqrXDoZdLBniZuXyUtiO8qYTUYlUTtyErbKU3UGI7tRIPZyQChwCQCqtE+Hii2uBZfKi5E+8q7P/nx4atbZRuKhu013ROBCKIE3xIw7IXdHX3nqyLTBONKjnzm2XT87FNplpkI77x/Pq0y08oMEqSCKjbO/Rd+6x8gTVgXhA78zj+/hlpBqdgcLbwzpqh43fwih6OKvPwPo18kjBiFCLqfk+XuogKZgMQfMXXP3KD+1nyPuKhnnfJ7iBNUKnWi1JthR8UCnflJ8TxULpFP1SZ7nbSTdsg+GIWIGCsCDFrM5r4bXrt4ww9TN9zR2ByuuWI+mkG/DZZys1oyAEdxPth8z427IJ6E6tzL/WPTfNeZZpj5hJ3KxyPxaLxie2yXr0r4+btO7B0bbdzcdD0WnqDOvlxlWqIRgsBP/MIvsobkFNnGi+nii3+Vzv4yM+0j8atMnEdzccuzRB6pLgeY6qg+w+yRnNd6lUbcc1Jy0misqyEB2AenuQ3xV+KBsYT30m9HJfP9XLGf7O+jQyQf1w57IbaMbIFwqFHlNTthEgI4eOpUmjx0MF1++dV06Yc/THcufJAWAWwf3hUX4HSmxx1iMHO4jxdu3Ay/fYulEWJEIp3qAraWtbmKy1M3J0QjwcSyDESOnQVRtSyrgLtkvd5kxhNUCjk+KKPK4ETYos8Dgi7Vo1xbOatxkchVgpLklhrQepeUkko8pVcbw9rYhBFnc9XM+DWwGKMfQZzwTukEYCsRNiSf9YEpWK6SRClgvQNnyEpQcu1tnIcKBii7CQYOs4CP4/zHD5Maf+1yqHGsQEQ5GnLCW/hjuxQcIKRcKLJy9wIprQzPRGaB10Rcvj955mysntVkksGZixejCkMs5zBz6QPSVQ4y8/0Jys7L4rmUtwvJ+j3bvL40T5rMNGUhlYhrKRXDbqP8LrSAgDWfDQlLfeyfso7xIevxU9geLSIB0EEIAXCB/vG/RU5dnSLy5OnTaeTgoVRHVF959ZV0mbEPi8wJbE5Rs9nPHMCM1WaAkcG7uevX0u233yGweD8mjOhAHbpmQ5wDTOZuB5kYBYno99u9zBZvynrsjOsOvZrUi0VcluADL4BEipzgrXB81R5xquwbuR0dFRIQidOBu2pLxHM8ZB9G8qBZjfr+8a6pg3vDQFydoGl8GyN8gyCnK94qXeK9gA2VUOqFmmbE2+70m/lgXCbSOyRmCCOrhBmm2zRAMGtU92HHDYxOso+nhTnmMGbOYe/tlVWee8kG8q6SeMfZX/Qe7goDr4OkwgAC0dAfIvDr+oujSBLnRnNWS+E7fIQ5C2BivmN8xJnyzTyggdFPvcw0o3TfIYeowSpkziXWNtubtps8GTGcKoyJhkbVbJ9aB79sY0gyjrEJrJ9we7SI5KONERMBgN6pfMy/5by95Av1MeX/4c8+D35tp3kI4+0X/zLUpanT59LE0eNp9gazOaJyzJNWf//ShZgXa54VaTfgUOroqk92rIaymBcIpezvKOZ38NoYOGMyipGhtHBvITVZq9BkvfTaSzFL5DbqmoQq8biFN6YwWnW9iixu9ld0HOe6mfPGdwObucc3HSrr2HcfVlVqoLerHq4RG1L67WDclv1evg9V8CwuX9VF4UR5WQvJBGsKiDGibdJFlCRhOAchgeA9SDFq0E3uVa3OMhMEZueYN0zkzFJHJLP+wEWB4ZyQfE8p4uq+IqNRLdcv6XSyazhgqCRj0xV97ItfTdNPnqXNnfTq7/0B36JtuH018p3kY32eaWlp8/jxM9TTtSw7EbTs6WUWFlStPH8zg+k6TFnUqKarSH/7KNeJLGPOrKEMQKh26QYNMFDnnwJxUGRsjwaR7O99zqPD7QyN59i1EQqOCfLpIRk5fCRciU3WF7z5zjvpg+9+nxkUrwaRLFRuphuv/wCPY1ab7n7wbnAhF/50etCdbfUrOpkyK0gbt0AY1JXm6ESM4Xa9k6FJco0owyk8jeLfu3Aee+dtPGdM8gB311WtirC9BZfWpiFnyfiAPWY2gEiRubdu1kzo/ubT7JCIp/HTC7wGgtVBorgPMsxdvRY2gmkecZ9yAzlBZFfqisBhEHjBvSF6cSQjskFEOLx1QTXRw6eB3i2Cc705PMA15wFgae4Dh6OdTqFa758kdYdpmmjH7o7pN6Ifm98OApAIJGb/e09Co31KVPrMNri5pN39C+8zpr4feA7AcPpyW6mgUkHpIIY72fkAqp7qVR2jf2udCcErSxALiZ24rhsY/DpX+kangym2qHusJw9xKVECvhKOTM49GCvXgUVR86jPXjvyr/9Pfx8+kYglbhyjwSE5gCTHUBEAisicz/XLV1MvakEfy5G5KOgiYx5un/+AKUrfiPl2K+jtjuvYaTMOgo72PaVCXrsEgxhDXOzMxjnlyR1FILxCNTwuupHt0AYTJLh6U0Y4PTRVxtXD6VB7nEnEctvUJbxL6sdwS6fxCekkBkUPqRZCqJQX3DkownbmJudOlDzBFzo6Uttx10a+FGpXm7ZEeQUydBcxhzqZtS53py0il3ayBqPmctlc36z+SNg6AFDag0C0q4SDu3CkWLg6KSW0RSnTQyqJs6b0mjqCA2BrPRNnWd9olC8FxRSNyF1HO/GMUT/VRYdRayeszdyLycuHjx5JYyeOIRHJ8kUiqjY5MHtbVZbBXC5T0eV6l0gXGhFlSIAyS7MolKyqgt3U3xWCu+uuwYINQl1iPDz1l0CALO+jrkq01qsklKgzF/6G28MnEite9oIEEuc0V87EHkc44F7sgM7thziaGJv14fF07YffTVfeeCNdY/jp9gZG+O5qcMXB8fG0DkJvrYAozosFNwsuDsBqoVLIWeX8pn+r+zdimQfHgTsmpYFXLOauonrCuBeOZro4viCm45kCIVgCeoWl40BkJ6czY3ZrS49QJgw71Pq7nPQAy8o5bHdvo7ysDhhnUJcxAxYvGjGC8FyB7D+mLgAbiVgCGTl6MhYxkoBaJgSuSqy0k7EcNdoDxlJXuPTKfES4RdgqNpV5U314lEy+bCFRnUzDia+diqkOfAbGptCTYEJIKQOEm6TKPyAFmlYQR3Bs9broK5CTYw9qZ30ASYDd1G4tkzPHwCy+MXvhQjg2Tv/iN7BDyOWaY/40YGW2sykqVd5bvHUzpKLxE9WqyFKmvXrpZi5c5H0YwBM0yFwyGQXB0u6w3QppQp3hT8BMlRAcom4SWEia6D9u/gTbwycSACyQI6Jc2h6KTDpb41XicP0QVYv68HAQx5HnvxgS4N758+lH//JfYRAyPanpDkgIkX1k6mA6+4UvpSvvvhmzeayhrmyzalUFIBuNrlQzJ5Xz1SPbVlUuz1vVN5wXI5UXlQgiLvSPDccVhAb2EIjGshBy/Abq2fLdG8yCf59ZHF2wR48V3HAH4iG71Zkat0jykxuKZm57BABxZIRDNaBjd8g164fLi/wdEHvDTFm+ahBR5HAZt4Gpw2ns1JnIT3PwlKP+DMK1meC749JwEG0bj52I1Wa2eg1zmYEOg0VsjhWQtEEcSe+cMAt3L8mYPQyIMuu5AbH06EbGFZ6JICocfSSNuAWzyafx1za0W655spmmnvoCKSi3UNlIOMWNvI67OufPddLBZ8/hbRxP11ksyHQd4yYdGFsXRK0avHTrRlq8fT2IxLSiMSTQyPFpQODcaK1Ys75SPU17qunWK99L66a6kBCps8A6aZJEnp21EmbUK9zCdqA7cP2bbA+XSKh4AHzfUU4sF5DDy4k9qhpU0Vl7GegzfPQ4CNeM1WZdB8T5tUbwljRHBtMNvFty5k3cqIszrHolt6FnJQANwx1G2HVrlNdUjbKKFXo65WuwNhiz7chHEx3BzAInrKM0Cyen43oZ8RjDecMDRj2FusAHqc2GpTfpZGIiqB8S+A6E7ziW7rArshafO0rkt3r88VvBmXfTMAhaAwZV9kttCI9ybYOOBhFJj9Xa3AidjluU7zn7vXsbRO+QXFi6mcMtjSqW2hnGeqJEZlWrGnXTBtC54NoplpMn0oC4QXSJ1Db82BaNzVf3neYmUE8gD6G1KNuhDHjqOBc2qpEuaLoyySTbuHvHnziRlm7cib6yj7ogANUwJcHg1CRHGA2Iv8Yw35gJx+HWqMOm9tj3LhQ7euxkSA/jOuWANj1c5njFFgKadoT8oy3C2I4sN+H+n7g9PCIpKmxVJZScv5TVK431HIHOBBJGMRxu8NAR5qT9HO3dxai9wmpWP2IZ5mOMfHP99jrxi9shztdAmPOvvlwgu1F11KI2nYAYaKvKgBx6pCQ+CVKkdwmB/nH0fDo7uLv6inWM3YMzp5PeMtQfhKsxqSRy/UZnWanUSCdnkJXxkjoEN0pHKv1a2AqLIInp9paV22vvcO6P2Gm///jcNPbOCOWOgMg3llib3bmDVfPg+qv3bjH+HA6NDaJN4gtm3W6S3q9TQmQMd7aEZeHs6v8ikXW1/O5uPHTUWbtDybGlSkrZbaPmwEaVz+fDhrKaxSZMfP+TtvguMFudvYsH0AVeB5EiDDtAtbIOq6xBPwdhjp0+lU58/Yvpjd/9IwgS4sQm2umsRv2d7ebYl77AdQKaZHpf+QHOFxhlY3gTFfNQgMrs474x7MbnX4iqxMA2iEwVK/hMMJUM1rgAfSBjIEjgUTYgAP9JLfnx6w8vLcUKs2eJAb3DYcKALlQsPVgO4VVdqMA9xs5+NvUfOApi96b3//zbGHS3eDele8zasc5IPyXIF//hb9IhzHTOoj6uLxheF7iS4l6XpIgCRWTDFY4qYTaH+5kqx1jKqG6gQBQ7XIQJo7kU0dRV472HiSPsRD08pmTo3VpkXZSlm1chktk0zDeexDX9zeNPpIvo/C3e34ULyrF7UAWNujvriDlOpq9H+ot1gTEIj8Wiww9QzjlUuRr1XaU+fE2ygtCZoI9BYyt3brDj4p65g+3hEFgJJEtOnxO8brH8NgQ1fvhkGj10PI2Ru+bQ217Uqcg0gPhNC9Fwb0Kg6vFrqErGldaWMKiLgjwGJ47fUvaDjVpRNzYutwlCOohNNXOENCDrG8OYlYQMeHPQW4sFmM788i+S7PhZ4lwscsoaji49d+3ll8Ie0v4bxDHjysCjx2SCeLbGmJ+YtHqlVE68rCNplqKe9oENzozWU8+zJOSMirFn8OZ6llWPtpQ/Pvn4cCSJlbMhxR4gplHRsELFknjkInLjHuIF2g62UE5lIMo1QeSEDhlVX3XQ1RZrvzv5gu5HILmnquRzRC9qV0/Mho7BCPG52KaAF0HkupkoQEdZktyG+mVbSYB7zUt5ImzHZ28RBXbV2zWCk45o1GOmJGhCDDfl7FXdzKR8QIyuE29ukraJ3zPabmqKaRqOPlQl87PmMy2jglzFvnge1yholnrhnm/cu8n4CmISUTeqY98HR7Re7MUWbmEQX9gKH8Qp56ysBaOo85znEQjk+w5Gs1HGLTbJBtjAptpVugDPiG7zkXAdq45Sh1K1MSdLwvC/BFlUpqwCahYrZ8G9WzAV+1ZV2SX33Byy7LJ0S7duYecRwGQa26s/+C6er2swt9l0641X+Q5GP6ttaZesEoEPLx3wiTw7mIeeN+3LTZwWpu3LUKM6tG+3CoMrGFt4t2yzN0OS8Bju+wjg+oxwA0774bfXiH0nnx6RWJly49xOFHndJYbSDnEF2VCD6BiNUtf6iMVFm6gHAGRZDsqaJA7gUX2qwtmdMMDhuGtzAI0x1WUiIQIYAPEJuLGuUuMCRnCdVM0phkRcB/oA1SAQEx55MBOJdS3qKfF0RVRZAOsA6k7r6NB6omYunY9Ew244ZQPCO4RR340qdw0DeptOroUu3ZOGXSQVQtezpPdGXJerzqcr8e2EZzqm4qETVyjrKtLhK3D9o/XRdAT35z2QbBG1aA1k39LWiLoV/Rx1zX+c5jRS2CGO9Z2lDEue1dsUDgDOg+R1GSN97QcRuMV7a6wBD8YxdSyeMlQuOiXGpOuadfLsGJNP7GIDNTCGIgCMDGMaE0Ef6wA+QnQxDJjzSNuHSLKtA2OASBw2PXfpckj9PqZNvf7Kj8JLaJrQ9dd+yLMsMQ4jFFbLd5jSFobonAASrPEVXdjii8uTu9xfk98CNEYv0n8xgE/4SAQSjpWKPzSpuLZ33XtuXv+E7dMjEitAxWOTOGxEEImqRkEkqFrBCQGG8YVepuI0bjGGSE5khC5cvpQufucvYoinNovxEBcTPfTZzzH589n0/h9/OziSbk3tmEpBfKFfc64LVW/UxAnWOaEDenHPtuYwIPGESCBZny84jNVVBZIPfQSALlXQGMJvXyGdAp16A7WkhlpxCpf0JPWdhWhnidPUcAQ0MDoHSU489/d/C5vqMAuhHkiLVy/H0tp6ft74P38nZo3cZa6rsIUkWJAT2ZL+9Pw76Uvkpn35+Ok0PjqWXrxyKb1v7GGLHDCfK/pVT1EmGmAKLHX1hgs3xpGrbmoXOBY9z5lF6kAmTBBKT5ZJlpuO4gSTJk4/hSsblS5Gf1bT8LHTpJEcwyY4DYJiAyEhXRb72g++B6EwRl+3q8wouhaElPrZhOUW60Z2oQk4W31jfCy+HzPEIDnvvvNWuv/+u+mD//htVEZSVOgDG6Q7/doPXqSE3JaTX/lmRO0PP/cc0hoXP3aX9srM5YvYoDfSElLpuV//u0Hw+cMwMiqjuuymJhD9yu9dxqBEsBPYBXPme6E9+KANKAHq733bz94m8eMFUfjdkBhcCwJR9APEIBIkR9gg6unsBuAOPPVMGqKD+pjY7P5FV9W9lOaJqufcJwvritwg06xbqFtOsrbpmGq4oGqR6o82iLZODcPfzpILGQtRfyXhPKSAHDUARMfGFvUr6kjdrZ9EaZ0V3s6vGn2qaseVGt87RJlfxzV7HS55n5ywdb7fj8t66NDBNHGWNBlymBrq1NgjG04wTVly5FmQxY73+0bIY6y4MAM2EwRDxblZVLlJjFVJdgjpugoSt3k+VCdVS56P6L51LXYuhmtYyaghXgUG7uEE0TYrNm0QnohAXo0sg35sAb+tCjNw4DAu50ORw9bQVQ48DbQO4oqW8QiTdaShMAh+TbXdovoe4xd/qINerHDp+559wu4DIm3/2IH4nnB2Mo2R6UO4f4+nEeZac/0UJxBUTZUAVWtX5mBK9KdTGWm7tFF9KSkkqEcLDk3F7+9VQhBH7+X73nOzsvuP+deH/v7sJEn5cT63V+GiMgIjak+HBkXzOwjFo8QSBENAD6Rqihx4njRWXXNdUZ1HxPEsnelKsWuzc2nxBtN0EgPQmJaziWwCwM4MqYKdkCPNIDhwFHG2NwsvkPUSuDy/B1NPSqQryorH+OMtVTptiz7sGtPBp2jKcaTU69gYSxCdKpf5V64gPHrqCToaQxlizi+jWsnVIQ4h0Yu0UZVwibgW78fkcmj1DlueQb2ZWVtNR0jDGcHobwKbFerewsWty1kiUhJa/wxnuCjEruHcC0PobFQhKO7CPSNmgyTYABbOPGkbkEcheVRMXcRHIqsRxAOAoZbmsezUEmSWeRmHGmDWE/PWJEwn5mhjCz7YhA4fLDe+q40gN3eR1YjM05/mqulo2IYoDM4aL9FTtzrbidkeR6YPw8wGQvVTKhi01daMWSBROQeQ2mZ5W997776BQY+dggot4XZIVDWlXrVLtiYjsNNDDSvqperpu6ra9vtHtYWy+h5/dpLEClhRj24cgzhA7A8RRCE57AAliR6XCp3bIE2hQVTdxLvb77yN5+PlCDZpZIr4TXTkAaY/dZDQzIUPWB7gTaSC4ydsvJPSgQR0upHvHsZK6DHpJVmxHx24fxiCQXUIcSCQ9nWq9Y36xVGOVxButEME8HGuE3dgtHCaYGK1X4fzfRU16jLu2df0DDGs1ADfoc98Nh1AFTzxt34p7BfHueumVe1Zw65aZ+/CVhh94sk0fob9ySfJS5qLDu2Q3r4Eoi+BRPiu0uu4f8dpyzng8t9861fSJRIflyhrjfeVmg8YkR2eR/y1sNtEKpsY41corwUnXmC5hOXZeyyKhOsYNbPl9Epk4To/MKDLqgpSQ9f21iori+EpdJLsWNwH165p75MnIHxsDRf1WZ9zbgHUG+AhhPaxmgBXSBoqoQGuWmtsK3LIQHql58RJMrmPnEjjMJMtgr7TZ84xBojhD6eOx3LfGywaO0yAeOHmrWCQU6fOwnwOwkQnqMs4km2aIPNkpCv1k3kcq51BFJEMKYOzLz2WW0kUXPLqAwkTP8qn9o4/XUlSVsQje1SsOObEM9QrgK+0COIxHgLHjd9KDwkEwDtz48TTqCejwxjjc8RDXoJ7TcRusGptZj585YraSy9+B6/IQhh7uZckTjxbiHMDT1VcnTWNf7h1bz9Tpw6Q0hBEIbejniK/XMajBLyvDUb6va56IOJl7xfAx7BWRetGatVAvE2CXndAku+jW6+y/FmX40Ag+j7qbKqF7/XDLV3GrgURrBer9PagLgz/0t9JDbw8uoKdhmjuvXeD60Z6B2VXqY/LvD2L6oNlkX50m+RNfl+hrHUNbyQQmBdtsFeDI/I9jwYZbapqihxW20SK6ShFkWpy2yb13OC7efojBgssEE8BcYVLZ305pKUS0H4YPX4CJpYTONfmF3mulo5+/gt4q66FC9ccMmHpFnCMcxUxrwFj60IdNql7rZ9ZH+nnxvBkGidrO6+rsssQ6wNpliENdy+cZwDW2ZgjzX6aJXA8CTEZf1Ft2zSdn7YrmWSaqtvGeHqRMM2Jo3hEJ9Py7eupizbgtgNylGJ97F/2kCqoy8IpQgXAJeru/aINtsPtp0MkFuxWHEviiMqIfOyqRsGlJRI7QQLhqASBnQQilpH1BkZqk70FMJeZFNsJlsdPnQr1xdTv3Z2L8akNZtRYBuEcxqpuG35yVAiBZ8eaF6Quq+u1p4nLlwFUeqZ4mN36ukscAo//1D/iMnSC78fE2gVBO+ewiGscRmmAiEtVjUF+b+FelHA2qENiVviYvV5CQZd2wjxHDFrh4PDaE85eCOE5XkRD3rR/KyQx6XWrwThqhRNjFEQa5bkpVM87TLC3COLfv32LeApL2Kky2KfREBvkhnohkVAfvX6x+RBNjsFJth2C3mGJBLl/GPUQTdhl/Da9xUKDacAEGoypkbHJ5Mr+s0yNduHlwDRHc9Z6ya6GAPzu3sb9jG8Z2EHAnALxiBcZ0DReFIzTbwA/Y1oruNTNojbNSKIpVUW1DH5Em3UTW07gD9eciUWPZ3NkinrKhEFt0l2q2FnildMTySD2VC9wLiDmVC54FPXO2Z649jMhEqHCB/IhsA1CEPnyHgCnUgFkkc5zd6WHkgSE8bdR8P6D5CcdOhwc+M3f/7+JrMMN4IJTZ5+MuYBtiINxbr7yGrooHhaCV/FpPwuglUYGDnsIoKliSSD6251RJVafAlD5eV4QGL5nnbxIHfoR6y6uOchiN5MQppyTO+md//jd1GHmxa3OCkvQXcJYR6VCz94EKdYxpgf5xgkM2rVu+D3xgTrpKxap5NNV7Le0Cby4CcH108Ye6ieXN+YjwrpysKkvddSRXgi9gZH8NTx7Xz99Jr1CxPothrhekusvzFJcdGcglmXnEXmZ0GQkwTgoM9rFN500Ir/jezItaIVrS0g/A5yqbKopm3ik5KwytfYqfcJzSjntw1odvR/ObZlt1CKZlONdmkye7YhK4zIdouf7N5/dRTUNKgWSInqdBMoDqFMSt+7m5dvXsIMG4/edd98Nye/Cpoee+WIaIk1FArj0w++luRtXULORvqhXUBHSB7UczcAshBZ12UD1GxobRRKRS4cLXWY2yiR4OxDUCgxDYpMfupWzVwbssI9E1bil80bgFPD12Z9ckigp3Ci47JAgCjkPSCfifZQ4MmF4HaQGicNQ57kq/vhdPEerMwvpvT/955G8pvE5cpSFRW/ehijei6Wp1yLKzBxOGHuOjOND/HehG9Ko6cQYbguH07B0Xb8BkhOr9nb+vwcAibNvlBQHAoBy9rmbTJJG2kZPHzOnnzyVFu7MwZ3IzGWfu4a7EVVnbe5+Wrl1PdVpbwPi3fl/2nuzH7+OLM/v5r6SmczkvovaSqVSbd015ZkejLurp20MMDYGgxnYDwYGth/9YsD+E+xnAwYMPxgYeANsDzxudHnsstHdU11dS9eilkoLtVDiziQpJskkk7mv/ny+ceOXP1KUuqgSSZVRl/zl3eLGcuJsceLECYjxFuOHcdTE3z9yrLk0ezHhe5w2UHI4kabpNwNPOtK5nb2MU+SABmpwC2iJUiJZwDARPRoJpZr1jePPNqfQuacYW/3rV19rbjCznHi5cP7wPHo8xJGySu/H10tpwVE8DnjOvXXxiEQHXjpMJviFKgtt930Gu0mjwQNJTx01JQ9jmPAY2s1uvcxLDMAE1vUjY1DmuMFJwDX6IuOtdtKQDPONhJkfNRbOE1isDr74crMbZJ+/OdvMEbTu7swFKTGfOGczhrVr98FjzTFUuTxHUo+y1PdH/91/3dzcxruhnXOahDn1gX9Hv/5NxiiHYvn84T//b6O2utzZyc/f/ff/Q2bsD2f8ev31V1NvHSE1JIgQqSbEI4VkklF8zrtUP39+PSKphGFWXlsAZ3onBBACkUj4SQgd6SF3BijeD9AJ6pmuVNt15Eg4inX+6P0PoutrOtRqtQA3dtOeu1cvZXvqrJ2gZUqO/Fir7eSjxOL8iZ08jJVmCJWnH2QWIGEVnERcrUlabkZxIcmmPlwPMcC3zm7ZIGHENQPuPXf5cnObuQ09WzOzThvjXAg3W4cDroFgC8wzrOKO0YMpaYPwrOrNK6iJumLUSTjb7MScKxE9CxO5ryZkz4ZYVa92ok5L1y3cbc4haTaoyzbP+ghWjX2sWQNAw6iRqg6rdGhM4sFJEJK8bacqDAooqQuSBhu4lZFFgkscGjbsK57J1f1332Fa3juekcSEawn6xyw9ksPkTu65QEuzukSwfOsjihQBd46SP3BXwmuFov+1UkpUlu/8h0xCK52qaoiae4krqhil+90wMFuedwKXxVwwD9XfJcaja2gWBhY3OJ4eEJkIpUVKrTtXL6e8YQhcohdXJJAwcPLvibrFKDUz8sICeNHuQj2lDZ+dSMyIw76pEmRHb4VYVKX40cpCHBKGBJPnO9e6nE8cPR4Oc+x3vxVrjIiouXcTMbuQ/SzeZexxlXe6oaACICHS2ZyVGi5TdZBOj6aMpEFyjLJfybDjAtuc2pa260Uqkg7rer9nf9aXuIZk37eONfMMABcR27chDBdYuQOUewbePHs2ABRRRlCl1u4yAJUwAPgS5d5DmlxnAdgyyLzRT2nAZ/7qVYwQTCYevY4EWMkE4ihrNjR1WraEeufi+cBJDnoXQlyCKJew5iyxMu/N2+eaDyCSEQbokxAx8iAm43kQaBdt1idsjjqqb5e5I8YjSCphU6VDJAjIR9eHCEQA+yUOmiB4hUz6UBiBOKolkTtyKymBPjRdiIT2g41xjHSwr2ST+J3AVHW8w/JoXUDiycs3HsEROsExom1W/bkN49MM7hzXumpawhKtN7sInE0GEIBjjAXytewyOXgA58grb91pzv7k580gli3Xr7gOZt/FmZj3Hc/orKkjanzW8K6Yu4I7D5a7PcdPshEU5Wvt47dDJI7fwEvbzb8eVIBSa2tejs9GJBKIP46oUlzHx0kiaCVE3EskDDqk/rSxD4qYhJaR+/Ci2ffCV+JaIKXPnD4NMl6IpeTulfOZFJTq7yGSXa/gHIiiskdVyg5mcDZMAAPnP3Qh32QSb2QXHIclverzmSOxbtTTAbwSQ6NAHyvw9h09wjr4480UFihnlxfhROfeOI0kG0zn3DxHMOe//CHXWohA1El8x0A2B39zcCewshmUCWAxuwc3usdA/hd3rja38AZe59cLl5ttLkFoBPHGEvTSP/33IEw8iEW4YA1SBdVlz6nnmis//Qmu4xebj954HfXDfTv4hoGrZS3x6wcGN1mbMQoxjgFf13ysdNbEy4zoXtLpWOig3RYnEr715dp/tiF9o9kVqRsTOXnFUlcRRIYmfJHC/nQpcXfdBQwoe5mJVzrLafUcjtMiyKal0LUpY3uxKu1lPHD2TMzGW1i6yJbM+E+mcS7l7KD85nkisvBtiJE6qho5gB/DfHsCFWsZmG2sbTav/cv/FSdGfLd476ZAxiZYQjojchgvFcudHsaXX/srJiQPh1EOsyhtdX4WyTbSHHrpJSQJ5u4bV5rrZ05jtj7O3BGGEeq7DW5GvRQuwFd128qWXQq47Do+G5G0GYSbC9UAt3DxSAsKrAumdEuQe3g2xMwg65jHJjB5+gyiWmGGfJnBYqxT5LWM28UGCHGAmLBrC0tw7V2MUW6gv6KSLCOSUXH0AypqFRwD4Do4H8J61dfvYiEsQxgAIrapZyxNqFcuuXULuSnmNHYhqpUyI4wH5maZK2DwrT+UyDTz1un4Dd08ey5ndXj1ctda65a9RKc5wA633h5i9hvpRvhOEXORjl+lAzdxf3ADocUeLT5wKscsP/9ZBr9uSrQMFxVBFm7gSo/kmH0PSTlzJSFaVyCQbE2A9CBRVBCWhGTrwE3ugRoD+9FmVeSKVc101Ny0tEHCysIj7kMgwVQ7DIwVYeHoShklT7gnxN7nJB+EIyNxLinISz597UBdGI8SDENYWo5qktw4E7JrTD4iRdAPw8XHCE8qM1vV3IzaJfz0lTvwpVdYcHUk396+cgnfrQ/TTtOYrxJkz1EnEEepKUHpcKnJikXms5R8o2gcTs4aPcbtGVQ1lYbOl41DoGisZLXFeIaB/sLtGENunnuPNEVNc1GZ0esNXAE3abZvFXWu4KvjIeCjBwU4KG13H5+NSMgombVnbmQX4ZKRLHIcGhbVqlWxehw7wJn6R7A6jRpggfdw3NnLuJhj3dE6MYV1SAJRAz6Ci8fyvSXSjsduvsIqu018ltAt0kHbw3AAOlsAK14doIv4UcWiHgA0Gi6xDtHxowBygg7ci5vLISSIRLmK5en6lRnEsAukUOPgMh+99x7xoT6IK77qg+Vrwx+DQ7kKbhUzo4HeLHsDhFuEew8MMY5BTOveoboTLg5IVkRYxhEOIId+8hOWHbMu//DhgGuDmXStSfOXQJgzZzIPsYALubt3icAialSltrc2ga8OfH38hOsqebp6sYxHCkF1pALPbZ/9vnOUG5GWjqKeEBR5WsZAD/0C/ETQCax7zvhbD/V3mUuYBMRd50hUHZXw+YGd2U7CcRWql752LhnQQ9cKyMxGcPo8+vW/xYThMcZ5zG/AMOcuXKCO5AODGoaxjePbZXAPxyj2pxJR952JY0eQ4mgfMDirrsp399J5pI8hVfFAwO3/4Jdfxgp6BQZ2FxzYi8rnPArbbb/zVnP05a8zaD/GqtZ9zfSpZ7NacwWmvIZECq7K4Pll6w/muKIRCZcdwD2idYuPc3jmZ4OseUy4Fhap0RIH3H0Kyj2ICfOFb/1e8+Ev30CMLsHNd2fl4PVzl9nB9W1i84I4xMbaCwBn3vplc+VtdnyFu67Mr6SDJJ4FZoa1xgzuwkluGJUC7t2PquCk+d5jBFsWwaPqUZ+uY4I8p46xW+8zp5pnXvoSaXRRJ8I5MaDmGVDfQ8WaPHWyufz6GxDHu80bf/wvUI+Y30Z12kadeeHvficSzw4581ffjxqxDYHQszFBYmltbiLltrDlD8IAFkGKdZAlbiLUrcypsIfgPbxsKWsAAnRxlz5USi1XD6pquN2bS3EdvCd2L8QpgkkoHoXLQfCqCbR7jkGvg9PyXncTZvHh5CViPd/SN6pWWvcyNmgJJowK7qt1EXFCMrx8YZMnWOqsT5nvVbOUJnos26bdIO44XH50SjVmmYjvcGn6BOoMkWysoA7ShhCdREHxEtcg0mWCyd6X/uCPcJp8MUQ9g3n3xgfvNzfOsAbo9o0g/ChGmy995/fpG40tGFjEK6SwlrPtjaXE6Jo6eQoH1mej/k28CxEjyS/+/KeM3VC96PtTE7/HDD0Wt8NG0UerQJPXW/wiS3wvv/V6cxCifgYGOcd0grP+bjmnqmvAu84mqJQrtCNdgWX38eiSxA7gZ2PKQL2MQzI52EoPicVZboMm3LqBq8YP/zLjDWd2h1B97BlD8ux99pnm5De+nmeu7Xjre/8qVgrJOGZHVQ6w2iAK7h8u13PvEUaOpSPR8Z3Zrs59Nsy6ab0amca0y+B8E2SfY5fZ4uDnQJVOoGM2cB3Z2t5o3vzud5vr77wNJzqf+QzLGmNwvZdFUxKoa8b1Kt5kINlL+T0gmWOrcHpy001mUZs8gNU5T+4aMywIqv4u0EV4haAcehUkVLcP4tseCCW7Q3GOBBHRBC9I425QZWKw+KIp2Rht4DVsWtUFxx9KLwhTVStl+Uzm1XJDniUdRNHHlnT2i2MNObPGDlXOQ5hQlbZa+7RSRS2zbUp14Kuerop25fXXUYdRN+HC1sV+skYr+JvJHDZpn8QayyIwPv6Nb0MwBHM4f6GZhfsb+WRBdRpzsWrxGIQ3fYJZfFXvMFwazmEfSjBOxq4wpzSP+47t0SDjVIDMZBRPBnFBDwY3SVX1UoVfgyHpSex24dMQlwz2zsxlrIQ/wljCfA5jG8czU0f2xXzsNuAZxMOQZPpbwOnB45GJpBJIGqX0sHHdZwqV49kRlI6euoEF50JzjQGdqoocfxAutRtfG1eeuRTTysk97t2iA0CMPgaXcjm5pQM8ASSns6M2IBz6DSKhgxns37dVQgDp6kE8VVFtjBBiN2oavH0NQkG6GUNriPoZ08k1Cxd/9rPm5rn3ucZ0CYIPY9I19M5u1I6Z0+xQy0DzHn5TKlN2QlnqO0pSVR13smWBEAQkYYTTy/1FVpFITs5zJcvGKvccNCvtCXqJwG36EF0Qz+8KIa2390mTvCAIa9JKkSA/hF7HIimAP5YUdAuBcMd/JYQSxGiJfcDF9SGu1xhn8m3vs88liLVxhJUijjn8Ld2AzQlsclMyuXrwLvNFjosqk7CdGzCSdWCRsK2cRXpDE+1GzXGPl4Vbs83199/H1Ms+kUgBVyvqJmRg8wn6KRpJrXzOlmeg8+HgjOUXPGDch+TW3d5dgN1K262yb108nyXJBr5zP0pjFJinE8ISrVL8xvkzzcRhpK3MCE1kYx3rIviQMQlpVbeEezex1io9MpHEhAgy1gzNNAVB5cUOrbVFdcLJMiaimKkd75+G2v86A3RnnUeWpzOL6hZhl1HDSo9CJFiBdHPf7mNG/fQbDMZZrwES5gexqGI547v7IGMatpy2Gt2Hddlz8rn4GO2j48ewSBklZAluNPPuB80gRDuBNWsWsXv2Rz9qPvyLv2CwfDGc2DUZew4/0xx5+Wt09kbzzp99L/tsgIGp3iTjpVhoqIOEk8iISg8GqGtLBXHjNiGwJVa+UqpYx1jkrCjXkQYSUQ65vEhc7n1X1CRh6LYJuK+I4bxP6CEyyBp89PiigjHeAV5QaNLJ1ZOVEi3fKXX5nnKV7iKmA93+0d2YRAnLBJLqin7kq19jjgNXeDUBqr+KCdofc+7RFuTgwnYX8Ny+a+RFFryR0DKUfvap28LJjPR3M2yq6tP1DzTdG6FlFh+s2fStbkEG7XgGK5bBKJQYaYvwSDs4Cz+e70Jdu/HhOYjrPVhUKUuYIn6b5779bbSO+fSFwQeXGfeIb0MYWNy6xAV1+57HS+MIsdlmLjE+wScOXDj80svNAZgzlJIJ3owxwa3iy8WYpNbB+rTHIxOJFBo1S8iLEAIwvzIWAYsTaEGEX2cQ2weQNoNoWImw7UPaIA+EMnmMxVLfBMHmmzkGr3OXLiAmF7Fy7KXjmIHG4XmDzrb7JT7t5tGXUQ8sNqpMh0oUxf3Fk/aUHqVHWJOAOGWh1saajnx9zcGXXwqx3Ll8tXnzT/6Y+Y9LIVqwzJaQISZPOOw19GUtaKuUN4p5dwi/pFHs7Ut4yLpUdGvTwaaWHRAVZA0Ci6jtoZoEqidKCihNXZUKmCJUHVNQqXvBZj4CqyuSeDZf0+q9GwQv9BOElFN7KJmiO7ff5nvTFQrxgv8O/NuPkUJ6IqiS6LI/jqvHxOFDkRpKZ+sX864qIh4Mqm/q7bdRkwYduIN49EImV5dQQWUQi5jGXY7gwjbVMesswfRi8XNJtP5oThQvaMJlHLCJmmO6ftSvrFOnTInBOgbx05YWRMDJqg9GtTUMFIwRZNcz2nqtM250UlOVS8ub8bzc6df5lnEmhPeQdhRv72nGu3dxZYkljrrNXZtF6pwLgx5lTJLxGvB21j4SpD2rLbWQo92geP7+qn9ESn6qQpEgXteMc3a2GnFOxVeW7iBmsSIAHNdXb64wqAKRQ2RUWLXH/fI2WcGnA6BR3t2PYhxbu41fWVyLtSkckE4ynI7IOcge4QW+LcZRd32+jB6y5+SzWeSkCjEygSmRlhpKp2+EzgHAa8QEnnnzzebyqz9H/cFtgc6zSXaeSJe18lg+HGcoIXoxH0Zvp56LS0QfZHCtQ6AqFD1betIOVaWCXgJo8nGdhu4i2QpOGFEPVSMPy+MVRx6WzrBHrKwnEY52loVgeVKeQ2HJI2lbAvCb+l3brenglGU97Oxydm5EQlElMajCbgLmac1auaOrO/EBUFuqeqhX9QqWogUmR8doi3NSVllJoXqjRF7BekXLopqmvcBALqwkVM2ZxmjSA1PUvL/ILPz89Us0vuj7GmG8TtWtK99VCWvzotJzVhq5ianjQFVCt8fWj2sVwpM4rJOLwHaxKE/VaxXVXouXbve7UOfHCfLhtnKZMoAB3LupjxfMWwMJ76Hq/HbRv7Zd7ps+JN/u49GIpH7ZZgj7KaoWWBtisVEU6MC3f2wu1qo53J5FuH5s7jY+NnogoTvBLJaO+WuEycHlRG6271kaRwc6pmBlLPMpzL6CUbojfPTBO9E1NRd2N0Tgus/F5MmTmbVX9TNMzYYAZVHVOiZjBBrhaX7RXHntF80Hf/b/AOj51EWrmDMPW4QbkkBuXX4/RGwzJZpluNYmnQl5M4a5Rzo9kFVvQDwVKjuaHMKxoZIt5byIIjzgPxSLGJdT8oxXQdqcKyBbKSJmcGj7tz1iTwixPK500D6XY7fSTxJQYvDv/qNwQoqiHvyDqTmzHUmiUWOSaCRIWp/fuXSlOfNnf5rPbYcThTyOlFD9knj0ajCncHzaZ4gmwxhtAFs6NlEhq4ex476Tr3ylOfHNrzfbu2B4U5jlMZFff+fNSHZhI96sY1BxS/K+1tiRRtpumhIiEU6k1T0/ljIIT1zI9uEM+JcgvMCadIde+jJrjk43GxC7XtRTSEnnxHRV6UE70HlyBMbs/owG3VAjUH0zrtcaElHJJ5woOJB8EJqPRCSl8kAwgCdTiCUICwWGCrGWLDMZtrKMr9PMlUgSv3FhjDsWObZww5cxTL5WRwLSYdDAAToqjmtxgdOlsoj0Neck8MtZwh9niLmM4llLue0hQMcY+I2x+s+oirdYc2AwMwnx0ttnMR8eCULdYS7knf/rT0CIMv7IYJ9v5XxGCQnyi2yUGVVSjM6BdMFdRHVRx8RsoS0YbS+z7EXlEx6Fi8ayw+w3yhadrUWIV+SlKiIcOp1vh+Qo3WE7PKJ6cCPix6W7pksC8/U7fm0aLvLNx3vWxvGOf5H6ELMD8Ljgg8ROvioIexnXje3d09zD+qdnwSK7GDtrn/kWCHHfCy/ErWaQOZmUTBusm0aPLJrqYT4FT4h1uLcqoISiJLl942bT+/75ZvgAriWMVWK+RzuIqkRfr/G9OyZr6RyfwFKo6mp9ZRC2ULxq+8C1Jws3bxGoHDWY/ompmcnKqSNMPPK95Um0k4dZZsx8ytyVc80v/89/yfcwNqSdQe8MOrjv5EkspPubuxhrFpmXG33uecanB5tRvBcOsDWEJmMNOSvUhY9Tj/rnkYgkH9XOtsvMzHso3vUX+kMZ80oAGOrSOQFnqvAxlg4AAEAASURBVMen9jEAOxMg9hE3S9MfKASwSkA1sDRcwfSRSBaUznC3Vpaq6nHKOg0nktJbvNYWb8fvgsNsgbB3r+IfBYdwRyjXOGysAHTGH0twPIMO3Dz7ASZN3Ljl9NQ5HJhOlStnsEZ5BbGKWlQAxVxG2/m60dhcDxFPBlHc0/MZT+1kCY0iRE4ThpiEkRLHdJTt89qIXJc/0kHSmI833Ycw9r59bj2S5sF0SVW+LW2s+bR9Rf22gKcMax5rn8i4AEJrEVpzph91xQlBkXgdM+3k8WeyNNqxwOwHH0RtEaFX4MYyDo0GmXVXwvKvEEkPEpkwp0wE9487H8OYDEOO6+J1NXKyVYdPTetZutBIJMrcWleuuK0GogTL45XLfY0t7EvHT2oe9VCFd4s5+2SO9UeaoiWeHphTYpHhHS68XL8zzDhpdDeMEYKbxKvjyAsvNXsgniEmU8UNyxPWO7V51DFJrZW5UKFgDT0mEHqRAAZavgtXcguEdcLTOBM6feIUk1FHm3n0WwdKBhRwleHqIvrtAlFNBC6iUODvVK1UUS6veHSCzz1CqpolbjgXokVmN/4+N86cpRPPQThE91gmeIEz20QRdP21wSHO/eSHxdeHjq2IJne3k5UDInXGFQCpvrfT+1iYpDhQXeobLaZQn5dJOqUJX4a45KJUivaVQTU2d28FN9+HoCgnbfV7sdyfR5parqNu5d76WJJlKbEkEudWRCY/1bIEzABEITxbYe6FOEMglp288klg7CSpBLKoKw4R452fuTtzDd3dfQ770n86+Onh61JeiWb5zg2YzfVm/S5LBuDoTliuwmzW0ApCJPSdFkSLiima862Ll2CKH0J8HzUHXv5qBvFOCC7PzWI1gyAlQlVZELOajUst2wq38HGcZzBvl2FPPfM8mofWQtxycJ+RSLwGAPlUKbkFTlCZME+3sJh0Lb5jS5gvzeIbLGuYvTXyLDIGHp/+W4kYuUvjBJqKZHeT+TJh2X08uiTxayvXHlo/xvDv1+X8wquvwaGuIBiWmmPfYAs2bJ9zrMG4c/0Gg15C7DBQ3HP0MGJSfxqsD6yXmL95M+7vd65cIMeSr/b8g6x37kWfNDbVtXfeBUU2UMXKAHKQwdxe3FYczC3cvMvAjTxZQefyUZfL2sFvfvePmxvvvxtVbcuZvEogANVI7MWEW8YXQUeBzU9VzUHh+J59IAKcFUJyVtwJLE28GglisxctqG4Il+960K/pBn4Fcd340/FK2drBpjn+KR0qxw0MgWMrI/Imb2U+6VHuIJQYFXzbSiGrGeRQmsUiYHLLZ4AtQUtIEG/pIhkP78nP+jv3kUE5SD558lnawzgFHX7+xjWQHwTkn4gk4bnaUCdGpYzf6UazAcK6t4he2rqWq/oIm55FxjHATcnkikFN5QMg3egu6ggBrfOtrinTz7yIZEKlYT7MseUA2sG2FkNgvHNQCyvvj1l3vXm3e5lNpy7obcCekKlzTEozTrL50Tyos2c1CydF7Z9Fxp2LFzAa8d0ozHS652gzMEW0fKSJ+Pd3/qP/mMlGvM/Bx43VrWbXyWfw4CYtwfHWXSTXdXw2IjEDGmHFJsh8N/5Qetde/OWbAB1vW4hBj1ud9TbhRsbHVbcvLsyubLPP8Z2is8KF6NiYIH3BodpidMRtAs3J3FWhtugcisvhzLAIHbVBl3o6WGlkXj296Mh0qnF5nd1dQwo5QA/3pvPlnn5b9d/CjQsHFjcdrGohmcD4cJvJM2fEydRKURcqQwdUVafUlleplXflKvl7jQSiRNbCM9BmgouMCtyQvOaXsn1Kw3QB0YFP02YARNrbl0Fe4JRBPNm1uee9EiZERz7Wzzwkkki0CqgAmrZxr3lbAhOxBeM2UgVzozUqDIRyhKHcW3XFjYOEn5YgTajIncxLOH7Th8w6Wn+XL/TxrI97l8xqABlUs2Dp7CAz/La7h0z1eNC3rofxxBDPhuXcFO4Gqelk6+FhvtTKfzKLcSxXxl+bgilmMpIxkzuKbSAJJLDMG7WQESbZHQwPAKV9H5uWGltgiHVF/eBjEIhsY3VUk4CJO8OeqI4MFwYYLgxCZEqaIrNLlX49IkHkHf3Wt5vdOAy64ObdP/9zOhQznA6HDA6NwCHXVAxvElVEG7pmYLvbiaC7s1AsBOLgSiIBQr6iU4gQgm65yq5Hcu7Jw/sx+8HhkEYiqvtWJAogTnIfnfmAzrLDsODAnfp0kaCjthDnG0T6kIv1IBlcu144X1mbHc5HJ8t1xKV6aHaU2+w/fLy5yRZzmwDTQxXH7zUn6tbtoXokYpUBfR7d9yfqB/WVIO2YjGXg+HHrp9zia6W0cT0G7iEvvsgqRdZkt5xx8eb3wzk1oSstJLoQaBASwgdeMojMc5AHF5Sx0xgMtFRSIkdFAQEkFrl2P/kbcbJXDs73Q7TZfHQt0YNAL2SJ4/Z5Zq9BdDm0yG1YJ9trlBQR2Wv7Ts+GmPfxhnadeSQJzNL4x6mPjoNIhTFXNg5Tp+k9mfVWetlPIZJa7cCJ/mzboXXLuGUv/oN/i/IcMyxmEd7r//0/b9awkDrGEfktxyzGp3GwZLwjP9N3bP+LGB9QtxwDbbHU2/aqEhtYYgCNZNehoyHGbfCylyUWQ0SCcV6u+/jMRCLFjWFROPjVV/CTwkMTgvmD//Q/aS7+1V81V/B4FUHCmdStQYxBzIGDiHYHVHbaIGbByWn8enQ6A/HkikXs0lQBhJVoAIAqtgeHmRQkhZHZJZJzP/kxkftwn2fQ6USQXqqajTcpS7cXJ6/mZi5yZs0Bx7rcFoROthCdKpDru1MuCOhzAVzUm61mHin0S9ZUa7nxOw8nMrVmmVCDQ6jZd/wiEdp0IrxIK7FpSFB1cfC4tUJnyrn58TT17ekbgRGgm9NxCxD3udfezHvL0w/M+ZpdbNvm0gIHw3Gh1/WedqbOIJQSQiT2gVUwKLeSNnNHTIYavkjifP7f/PsJzXRnZiYDYEOHkhGqBioY62niIyaH4qiE5rxQkI86u/mPSOq6/L0vvAKSrzCTzdqfWdb6MNfVrEA09I2mCWe+RcyeHgbT5BlXEKQ7+isrN51rKpJKTBZHPJM4ZXuWmP15vefEcea/TpDOnMr4V5Ow9YrHtXmmb8r3Q0is3ezYu+fUs80f/Gf/ORY71ESYrqrg5Z/+eTyIl9kY9vpfv4b66GKvdQbwp8iL/VvQSuY+/JDn1LXr+GxEQm+o62me02wGpma8cRUHuPkrjknK8tWDr3w1lClh3PwQ6xJiMv1A56hjihjIRAiIcKBTrnEo1p/CE9YwT7r7LWpCK/Jjp6fyhrnUfp8JPxBE7ify2mDnAnRtKZN+5CewObxPB5CQsFmUpdmxDATTkS2CqOr5rofdXZPezvM/55YOkp9/0i2+L1e8L0jmWcSXsyPLoxIYwUVGUq1kEihBQ3nv+IE8hGna4aDcfPgLt48PFde6ysQKZ134V9JQhJIZIih1tUjmqpBGI3DEEaw6SkoRcU7iAP5KdksIAXOtRI7BgGedgzqlScAHesylUlQktZyMzSjXlYEuo5XhCDMH8v1+y6HqY9oAzWfANZKDMtMO7yWO+kvefMg3tSa6isSgkrYDT7LjQSE61Ga9rIsqCjR8Z3nAQm+CySNHMt4amaK1MmpejzOBuoRL0ZJ9A7wXMOxc5fouOOtyhiX9whgCJE/S1+ORiMTOTxMocQvEXWWAc/PMmQwKjUwy84tfYNVi0g0EMEzpETx8D7z8lbitXPrpz5o51KP5s6pHqimMKRgYjxI5XT1yGH11S9fnAAQkBUFGmIHv57eMH5H6ZyxgAEIHN7d4c2++Tbk7IHA84oywri8C36DLBdxFFKuyCCm5qupRXEoU1QE8QPSZN7Is9gRxzkS91iMEEjB7V8ormeUtf8gz1ibveUMSuyUBL3xAfbJLLmZkEcx6KTVlCgZqcxzhz0FvPvab9tAnKjPEMZOWuhSCFRkhF5A1TAW4+HwIwhhFsk9gGh+F+Uggqr1X3n0/urzcV4SSWOToRpJ3lWI5bFupu/WQidhXwmZDeMQMLjcvpnq38DYAw91rlwoR8+0A5QVqJau0R+tgiIx3mnvtRwnG52F8AszDD22H1zyiVJAWSQsC2w4ZnoxQBqmXhn2U537TEiXyBZWL9h8icKGuK1jCkjt/Jo4eibOtdaHCzcIMgT1gHsZ6Mz/3gFlvd86yCvV4JCLJR1K3wxoCHmgrP/0//g8xvcolU1Eqqwewk3zjbGo/ii+WOvb4/r0ERZhr7kBAwmTy6LHm5O//YfPcH/zbSCHWH7Bs9of/1X/TLLO12jrLNIXYzffeS8cLODmESGAZAmALBzX9qxw4ZsDOu0163701VG/sWAkiJlTUKweI6RifY9otvk90lJxNAqFdlgHk7Z1wxWqLV41Ir5EulQ8gRPVioi1jizxMh2gSd8NMl5JmDxKkXdxtsLKMMtvtM8vVcubOVPdYx7+ME6YqZCRQEBMgUR+5oHAWHgWJvAQORLkvxOutiGuQhkFWAH4pZs6MByDytBOGs3j9SuqGOKN8TKK22YZ62C7uKkEIC9tsvmAh7zSKuBq09UsDwVZRSU03dey5uKL7XsKX8DTxqm41TVnlaLpIERlgfhKKsC/PC/MluYeMiaY7G2+UzFtYKB2f7MVZUeJcwSfwxum3wqBllB4Zv1hXAKTqtTiDpzLjoTWYqzEGouby2kG8SwWsg9qO5dPQTFoXoqVOtG2nj5P9I/pupUa2oADUgWSPgzIpU8yXmvlto6rEQkJlkpaK3GOj+0V0PkN9eqgrK9pcEWcHu6/InmeOQsmEp7kt4gYXgBets7PNG0BsM3HoGpLRhOFBjcAfK8QDQmHEKqv16FiRQw7hnuDZFxxLxzqmYE2YLnYit3zXz/MyMC5qifiSsUe4KOUKfP5HZNv0YKrN4kbE4n0fgHcw7syvM9YitxLDNRPOC+mjZigexw4OcEt/8h1IN4QlrZcBpjq/Pk/xQYJTOikr10zhKYlKBHs4BSalJlqU3H4gbu9YGN13Xr+4AND6iojUVaIAkHwKDIUt8IzZGIJ2XFIP61TUQoDJkclUCcq8ODu/MS/BAR8jpWjJ1BXFbb9VaaN2ybgiFfnO8tufCBk4pk7Uxzr5a9Uhm+Ve7KV+hQndw/nVeRvz0HJq5Jz58+dZ36OnRBkzBaC0y77SgKP1a/nmbDMLgU2/8GUIpayRsezUAUKINCNPC6sEonTxfSpQAcL50SVJPg7ERBEaRcN9ZmFKE86KM10QVlkBaGA2K28YShfICEQPnQuNZKjXqYewcgBu51VqLhYOIAfgtJl7b3oRUFdov7zHxFiA5B3I2dOLqiYCQFjRnUFefY9US+yEjAUkatMEIZAqtoF6d+rCtcCSAehdGkKxMMu3snwbYHLnzP8A66d1kTBkTm8/1hqrax2xWA2DwC5NHQxn5VvyKQgr7SsBjCqJZLE9lOu8gJHaRRwDQtRZfQeWBeHNwnpYP4lRF3gG94xB3G9wEEIl2wA0GxoltVXG1Olj3oXAvUCqKC03RAyOWOCEi+1UitEpmzIo1E/L4gHIRYgnzK+WOzgqBhipnyUNBKuD+3TKJnHS+01+3JdyWyQ0r/YnOeQwqRchnlS0WWfcuwXjuQbDUDV17coKktcjROGFdaaewRfydChgOKc5iGmUqPijwNf6O5uuY2uHKCzftrfEUeDb1sV82+PRiMQKkKlcuJ1SDkEEgFQyHMHKUujcO6ebRSJVDGCfVjdfYVCkeCSLAHIVgtG5TAc7He5GmOi58fbpZpXJRW30AjDImQ4zz8IFa2c5r1AJZXkezsuKNCcBXXlmDNxVv4fLFp28DdwAsMb27GeMI+GmFYxj8H4VLtj1B1CRJCLHK0FK1DYrK1JqCs1glKSaPKOOAPiJQ8cz+28m7ral+XCcGMZTeMGOwtWtbzqT+lRO1tHP7SABkgoAP45RGMUoMNt75AidzsIziMWybp07S7ucBV+jjgUZHedMHCGQGzq4UkxiT0jV5ESTlFr0h+/GjQqDSrfK+MT9iCCvpOrBcGIdVQGHCLohU7A6wl9jQdknnTEVDqrCMvDBs9dJSJ09lToGX1CKuKzWMeEEcxu7iCnwiUcHFimIfFTpyhEG0MI8FE37NxiXbFw4X/ChJqxpvBf3uPcIvCUI/P3O//mfok2w/+QzzzLmnW4+/H+/18ydeR8pBPMBVrU/VL/shzyjjSHm5Fb+PBqRWIk2MxsAIyqEAVQl/nBiK+uPXVDXMTNuLGBGpdIZeIsUHKKFaXqg+Lf/l/8JtwHUEDry7gcYAbSAtdJGgPk/q8ZMTz7mZSgYUVx79y5m2Efw8lybxsWCeZeBUXR+VSyeuW+Fg8NSOZA9BbuP4CgqlrPoqBk8DDGSo6K6EKblWEUKtwL+pVzVE/7E4iZCuf+7A+ASjALVTtMnC+8NiicBZ5yWr/lj4bQ/BIIkE+HkYH5fDitHG1Oml7adMrkUzFOH8FGzA9Mc4Ui9qVOCxQGvLdaa+yx58JW4zgepg4+nmPDd2jyfSVolWAavJAkXRur59TLMxVn48rG58aH9zc8t4mQX1ktrFqIvdXWcMDjKfAOuRwMwj91w7qHRSeoG8Qku25mjwLFcdl23b3NKX3AlrEiyjQQD4XLtn7SpTd+BkwkhktpX20j/1JExj208973voW7hgQ7DWMazwHXtxm+ONG8JRSIVrwVujDrddeL6kYkkHL5mksa0mfPMDkyLRGYAquRI5bs73iTtwZfNwoWzpZGoAyGQVNY37cG3aQDnUHjKVNngM5KVtc1wTKSE6k2sROiqcn31TvVWkbMc5kt+DO6YrAgSVOAK2LYBAXIhw/aznIIidokfhktbhkge8U3jDZRmxEgD3fXhYZsOtr7tYcfIrYLsQfjybZAiaURLS+Bnu62nd1zrm1SPDhx8QP6l/KLG8hXIxPcgjlpqjH/UUw8IvX91mY9VSQThX1QspEEG1MDL6CYeaT/fp/ZgZ8ziqZYPRcABxh3sqYLkHoJIZBiWpeqVYNVwGaVY2iE+2B77xLp5za9ep8DOnxZenEjJYd95xYPOM+47cPWdfcM5yTxbnrBmWwnMvJaXl4G/+Uj45IuUCj773PyEiedO3nzG8chEYgZkk/powrT93vjMN2m8iMOLAggrzMuuw+f1KH5Q5U4Ej6RKJUspycPXLbcwWohqRS9pN1nmK6DVq4cxFQ8fPxhpsjQPcTDeUMQ6I7yKa4oASLkUbYejxKTOkQ62iTqrXpSWlPp0/lIVTadKSsvTQ9ZjfRAXDLjUIoNE/dL2EhtqdO8RBrQQzxr+TFh80grhUwAUIik6MXXUwiOntcNqguTcEqRw4lfqvQOzJEl6rqx7+/O5qSJx+U6C7O2n3cBCdcp1E7rlzLOFmu31O6OF1KMSMAC1WN7zR6JxbIa5PeMaX3BolvVTCd6Vm47HVHfNTbwo80S6pUh4xSughO2ReCAW+5N/Nr37iArUeVDwqRt/Qgyd9+0FGkAHyaxeRUouQ4htspwqrIANbwMDauxlrtMf3em5fjQi6QYoH1ufZNoCrqj5BYihykCae89tmlSmTfJAXaxlgBzol6v2Lx9o9SAPB9NCVj0d85A1iMFgewu3AgjHGWrnB1Zx+V66R0wsOm6E2ExuvaZ6lU0v+VaC9Ii7h6VYv/aZnceLnRp4q4qBZc1xAl7baIREs+914kkuiis2ps9Z1lon+N40kVpQ+W6891oky/TJY5TPCrkO/Kiz9QdR/YVofJeOo6xOOqtF4fWXGnf9MR2/wv3a5yaXQfkzP5C4hx1pjUfm+n5Ns4ZvXWU8aNyugpS2n/QaIZCEezDfT+JdPUH0dsc1xtx99y9+ENNuxmKUGUYjHGEs+uURRRDmgM8WfRJmQt6rRGLsd32NdfG5A2i+tV6qP0r6EAxwpSHl58kr21YuylkY5Gift3eBD5KtwIiHvI6Knfct7Nq05WQC8/DMZ1wX+LXPfVfLbr97NCLpLszM6z3XVrZTGM9z37Yr7esu2MrVb7vOHwNMfZfEVv5+zuPjan62LpEyck1swdt6DDNwX19z0gxrG+Mfj6Rn286sNyBtjg6ylFtbFkIpt4UZpL0pJe0UWRzUOjeQPGmkM+YGvZYw+zHpuj21m5kOjRGzyi3TcCnp6Sn1MOsKLwlGZI0+LJwspoWuNcl/4evj1Ma/3Z2bF+UPyWu+QUAlFUgqIXqvBXGS8c0c99tYIHV1qUeQOTfkDtGsLJXJTWGXgNJlJJgUYY7qu2TvoF3jAIBPGc68C8NtLAS9I1rQgCZEknkn4u7GCiqDE6aOGUUQ210QpVxTStTsWrlPOBeY8JLvC0MxoU8L3AJX8+V/yrAcD8vuur7vvqTo/P1sRNKdeVuBFMgsej0i5ngXixCAzGFF22MnZX1CvVtOnsbUxwEcbWz126SxfBspRyJdOKb33PUOoGfTWUMYA4Z34WS5wkTa3UWuJ5o1iMI5EjzgqBdfohbZ2a75ygx8yhfY3FdA13p4tgy5VoE413BSEQIpItHoJrPAslJVDwNID+9G/UIVGZs8zNZrrougoD6taba+tKEMxpUqcmXybmFQO7BtftpmxSrcKgKVdPUpNfMDicIzP3m05y2QUj839z/Z//wLqIwgPpbEEECSkJ4DfsEMPwursCjNfcR6f6SILkhrWNWsd5ifRM11LF/USF+5fh0hoRNr4gTvNgxDFc/AD5EwSg4JFYmvZMuAXsagyhw1lzqTb+cIrHfaVWDWefuxC+vl12m/b21/PbfZVJial9em6DwzLc/y87rr6Hnp5d/pqknXm896WSvn993XD7vvfmYFCyo8rJ4lL/ITCHI8EYGHXCPG5UpyKhBAE6yi3LNzGOXMlmcuICJa/Dw2dtWNRJ4njUEnnCV2ADr74YeZBCthewR5IUKrWdoikgpaTMKWQZljmHkdvEeKYFruxbqmx4FWn2Nf+kpMx0ZrmSSa5ZBrKGBL8xdY5Yfqpz0/Jlu5ev2JKO0YpRBqSrel5aLzt6vbui5TT2ET9QkYUE/hoSk0qzkZuHutv5bBpy+9/hpMxCAbIr6ZC+MC23LvM1sNLGQsuVXyKRW5J+3o1H4mTPcxBsTSiGftPQJUr2JZdOHb/mdPsdfLSZYCv8g3m4lSc/2N15kxJ8AdRpW4gyjNLD9MgvGijCL14WwldipSyn/wbzeedV+brvv+Yfn8Cvl/NknyYCW77y20VqxW6sH7rvSCvZua+Xrn6P6+vfZ99EiAyFX7K1fZp7stKxHDIaAtEEaiGlBqjLEOZf80ncDsup+jvjnvYPki9dSJZ9CjXY/ChJVuF3A4pUVQgzSpW27yhwzsS3RyJqgySUUapYnI00+ZbrGgmdsc5i46X8FcDAaGAd5XYhNWtf3lLJFQUn5Wshwlfb3z3AWprsswEL/1PcgWZOa+SuAQgxycciX0YZB6Wf+o1llQb4NYfqiz/0qddvIrNeDePHOjs6Q+VK1ZlWcynFVWpqpFrDEOMmSsazUksiHGQv2E/VnH5d296DNWIV3KoU4Fd8jZ67SjlPgr/+3+5mF5dL//FTP9/InEgrsr8rCKdlWuIkjXo51L83nwe56FPHxuJyZ1+zeP7lcxMidAJzgPM2IcrXFntA2+zTgFF2qJRG4bIjl+gAk3PANYp+ACH3UPkTN1TF0ozKJiV03BUd90aQ+RSHk4VvIJYXCIEpk5IhdCbTXuuWg5+hQdOHmEBCaywuXolGE5MUxwDiXXFCav7eWZ6R52kKYMRqmL32tsyLlwf8vxvrSZWXOQt7fvFu1Az/Jd8rQOXChRUh+eW9VOmdx3rrFkQSQZh1gOz7UoGnLIeYr1xRMZs7maNO77SPN+xmlKt54NNAD6RjhEemG2zWSm7ezk/7BGPvDsYWnJY6e+v0L6B5J03z6+Laq7S3kM16XPOvy4IALAEn+FmTDqPmJ1odPjSq6KBFK4x56J5YIuplpARx9nmelu9k/Zf+rF7JJUTNR0fquClDxBEgjIWXnHJConCWpAXq7vcP9xLWpyyZiWUSuMpbUVJ0Air7tCL3U1nxaZuc+1Z35BcM8env+mX0nZNp4b2h9ZICBAxMqxq3oaNZRxxODuQ/EaGMN8PYDkXWI2XVIMOaZ865PsSt6hnrZeXCuhDOanG45xvVwzdA8nVfcSMY9nv/at5tjXfqcZZ02IZmMtXhuur9eNhYwzpgMGBadLO1N2bXdt11M8Px5J8rgalE4rmQeRgGxFqKpmhEvKidFv+wB+RULP0aFBGF3P9b518VR2yLKz/QexzN/UYZJ5DvVlpQLqFJmAJYUgC7FQB55ZNm94JydEghGDKrv+MiZxHUx1c1nHHO1a1UgSJvScAHUZMKPYfMuf0qgH/3a198FXH7uvaYNtvBWPOz/a5vuun7Goss/9CIwDa6DxATKeq/lYgNe2CzUp6poZ5j1nCMr2SByqp8Kz9x6bjOJpoGQR3q7qXME5cZF18qP48d2dIWwPC54Me5pomMBQ4tVL3DFSlSqB9CeAxGo96eM3i0gqdNrO87ZaeXYQoBBGHbfkvemD1HBUvlHF0ltXFaksG05GZGHYGgbTSIesY0hna81qOWukCXnJI0kb3R09xHFmWZHoehF/ejYb8giXdJClEBZ5gFjlHQ6NINB2DwQoUleE8PrzPGp2Vpk2CIZ6aKbu0wOZmfMNwssaqtY2pQptwrSUNss0tonCr2TqHCQUwWUKBsHW6zcholC1YkEUPrR9BcvXKkSkpNab1yUBK7oLOV4yR9WtqFwysaJ6pRj1O8oIcXcKfToXv5lEUmFlR3AdXd2O5Zdu9KHvfKQ6AIFo0i9dzIo0nN9WWOC1yhr4pdvX4qCo27chWtWjo2OvoQ6wZoUMUlpmn+X85iIm+VzCS2fzCG6b/Rtxve8b0qkTZ0nVfBwIhydZS8MYR1VMl3qNRM44b/FL3c3P3+dxtHBI4x/IL0sAAjEG0hgxRgiwMcDqwpvnT2P1u0Lk/YudaoS4+d5gChJ8YgVEmpZ6ZowCLJxdjykcqbsGvO7NDkF0OhC6XGKhmcNfaurWjeZLe1ktOXSKfthi/clMQuGak4ElDMbQi5d3ZumtcwsPu/GLcPxmEUlFpHoGguFE3Mc8GY4kwvlCBKzwdtTgM5CTk4uCHIc4TnEW3DXNGIhjlQly0JFGmnQ8UTJTu2CQmYw5kUvIMwhZJumiNoBMm8wVae1qjNKCylUJWIuPwfW8zzOz8dipZHne1vlzQZCPZRLABGZGO9x1YF+zm6WubwKP1RgzUC9td77D4sdkrIi+uep3RaJY3UhRTdaOs0gPVwmz2MIhceUeyyFcSyKEyGcZZnT13fean/7P/3t82gxva0TGhKsFvplHa/utMAz7CsLjbwc2XpdKefXEjy8ukdgbDztaJMsrkhQi4a4b2QC6z3dEeUtESUagAlzMFQoDuMeziUAGlCEa3it1OuvJIQo7zryL419bJyYUy5bGdCWdVxC/qApyUE2iPetF9Uq9yLeqDcWAUPJMvjV/zq0cJPWvcbT5pdy2up1yzNa2wLnFOaWA8ztux+ekIQ1vGUMtH2QlzGw+g2AgG9QuMuVjGUhtk49UmTwXUziqk4jPF45V3J/ko/fPYFlkvILaZdDrgT637iv9V4iDj80gmZkh9UwO/HnKxxeXSARMANZCqANA7tsO8Y0dF0T2mYNABpPx6eK6DAoV5foT+bPjmmxRx5CTybzN5txbb6J6GUl+KV6yS+zNuI3nSLx7o10xcGWmXi9XO07EsKPjkREPA6SN7i3krbXLpbZ9qCXbm7ift75JqtpR5zBDOzfhUQhLNHgMR8GxknHnmgtgmHIZfzg2cK/4G2c/QEUqE6xZ9typDmpqJCkPaLMOjkL7YUf/CGGQkJK9zEWtsEiq9AOOkXyv1F66VRwqdRMK8wIgw4yJetnQNHv6tfUq9aOUxwSWh9X9V3n29InkQUKote4CnI9CCDzLYM9v2l/Eda4hEjo/M+/a4J1596dTnWfYVoiEfl5KIG7WReBLdeyVb8bsaRywuavEYsIsalmbWqPozAzIzY/ttMmg1ENOyjP4KhyV8Q5EYt5KI82hIUhrqg4PgVhHESeTiQmrVIh1p6m0q20Phdja0r6a4Fc5+71Hzm0eMAYql7pV5BR+Thxe+PGPcAI1VCl7jbANtfuHSAyJmEg2jsFsh89oJASPNa5DJKWsUhTtwgM6YYwglEEmCoddzozqucwGo8tIDWfglwhxKlzdSk93nV2YnN3Qx/pat+ohnElgM+YnTPymSiyb9zSOp0MkAqEeXdeFEHxRgFTuy3X3s/JcpOcdAA6gq+SQU/Ms8wAishyOd7pk1LUmBtLbWlPf6G32Hj3e3EP9crOZu2z7lo4BORIC1OggLZKJK9ZBhHfJbQ7q7mSl68DjA6YKgvRwjX9m6qlfCARKiUkY5HFhVurf1tu6dw4v23svq2rTef+wi/u+L0hV8ijIJ4zCHEgnoQYhaYyhoJbh+onlhcqlc6UqpUH8Qhe0NTVLu63W/YRtGbZDODsv5JIBJ0oNVj6gcYK6GsfAcU3/gI6UxfolnIy64grHASWJu4B2tcFS2yJ3nvu+dMDDIPDYnz0dInmgWUEan4E4hQDkWSJ/IQB7q5oH7fRsjyCB2FF+468lkvhtyfnh7vHd8lpiMVgB8yO60O9/7ssoz0gAiGSKnWWRATBL1z2U2F/q6qpHWRvfIoez5vZeJBkdT+Gpk4GWXbXnrl6qV6oovfy8LmZhEIFDJDIEj7GKtwiaZztrm2xH5xBDuu+9/iQEqek4m18+lfiARWDXDZeWQBJulSx1n3FuJFstgLh+b0AHLVm2MYiqlLSdVEkPa+tRa6pkzqw9zMfokDqQSiBuZ2CAEMcm8/3Xmj2EizXvO0iTO+szkUiZV2Gt+oCGDCPPp8k159zdD4OkeHp/niyRAKwc7VngdQiDFyK/9yKJqk4Z2RUEkAh8nnd2fpuucsdCJBJHOwaBwxk82bCXew4da6YPE4mFMcgCUUh62WxnBK43jERwJ67ZS+fYBekjIo2zeIrnDeFYxZKUwbmIe+tl8ZAvnNCxhbPsrhjULaXOwCt5JJAVXOb7ITo9ggeZYDM/51Dcjm2NmXfiCaUtInM3HCI9SkHl/cMIpIWfsMi3FS6eKTOwAj5lnNAyCVVOmYXqJ2qpWxUoRaBnIq2w05N5UcmyuzFqJOXaDplHJhJpd4dC7ESJBKljqNBxHBn3IZH3sP5kiTX0RlxxwvSZb/9hMzU1zaTuOpswvdWcP/1GM3v5fHPz0vk4QA4RSqrpJcpjW/+dtgRLUhylPvXjyRGJgPCwM0Twet0iv11UET+I4/NWVajPPee6c4aYRIoQVxm098CdDJAwTnACo9qLj7MXiMM0wPJSV4uzvmSY+RDd2+/hhHfhrdea2WuXm3nmTkZQw1QDdEXR/VvM9p8cVLVBFUz/I1WsQiSqcpSHepEJSBcfcRjXaogIKlA9dxIB5XKsEHlSL2TDtgbjhAW/EAZnYZPnJs61Fxzd1+39DkJVeAJT4SXcQigSiWMxVCylqc8gDuGnhHODUccdW8RPi7GDfFUR/dxzfqqaVksrRfEtSF2spWOLPhZalZCm+MRJLDCkCYijSEg8G7ZgNtShr2egmYaA7mFFM07vHaPYQ50SYg14F2JIYTbwi3U8eSKx/QCjdHJu7IXyzB4KwpfO/hhR+J0dZ8dXdcvOl2h4VsR/cdqbPHycwSKuECDlLWI39Q+yMyvIPAiyT8DNFwk4cQviePdnP0iQON0iRib2xlxpoDUnz8raEbmqW6ChimTwX+Y+JJJiNUOaMB/igF2ysl2mHXHpMOodZBwCEyHcTWqdQfPeE4dt+A4MghxdBBIemiSBCxUoN6bz8Gyb8533wk8i8dfCx7oKpxCIz6iH95z93s1bSYzEc3Wl6mYhjAzW1SxF4LZfuKNGzoi3ZXOnO08fFr/skYl0ym4CEMreI4egJ+Jz4UB65eLNbMjawxYLxgNzJzNjkA0yQUmBZCaRFAZSMueRRwry7RfjePxEko4EupztVBE6apXP7Wg7MZ3bIrpAM106lGvTeA03LsTQvgMZTCeQh+CKI8xqj4GYd69fRXW61bz7gx8TmsctrnFBQWq89af/CtMt1hxUo2XctPVXEunlqtMn2MecDv7owvmYeoP0rI1wZ2AlhLPCLhzqbyOJ9IdAVLdKhHhxuJcdvPoJBDFMUAQjqchp3aHLveYPsd3YLeYK9GNaYXvnYZYXDw9CTMLAtvKrnDTEnwgORcKIlxVZQhTiEOVW1bQDO2HI84704L7sU64EZECe8UaBmWbqu9evxQys1Uq3kRIPuBspKbVUzxIZ2NMPIjWEk/xY+enSZ61Xg3gYuB/l5R//RXPhvbey/Z/hT7/6j/4ZA3NUL8LYTg9PN+fefx/nzqnmq3/0D5vTf/mvmWi8i1cCamctqIVHamHD668CIC+e/J/HSyTdSFABwDkd65lOza8SQkswIYZwQKrnOzsc7mywhWFW/CnijTCvWuS9joVgRwJwL2veZVA6BiIaHSRIjmRw11UtK/2YdqeOHyVPCY+geRDMKoPMrG2gU5RGBnUD5Zh0HOd7BuMDTA7K+eSsnMmG64rchRuq4u3uI/4sA1+oL9Ycl7Xq4iIiRC1zBh5iy+Hg3/X6wYSaH9mTb44gCGgJBVbisMz8hEmYxg4MK8Ox/kV6ADPHH8KYZ35nXrbHwNn6r5mHG8EOsvuT21WsyDzIu+AsVjuIKV66EFKJsF9gIMMYBNl7jUUG7B1zbRtVnrYmfKjmYsYyI0aOYcmAQn+F4HI6VA6j0rqHoVuRD7D4K1Ky0+YC09S1wqVA46n+fXxE0jY8XJJrgV9074L0IlLtwEooCTAtN+Sd+q77gKvq6NGrE6G7WY2zq9XY3oMZdLoE1bHHKouGFm8TsY9geCu4euthOzw9nY6y43TA01XEDUUHsONPsVeeOnpYNGFS3XbO1YrWtdbT8gaUHiDK5ro7LLEkFSRTatiOeK1qBeM9HzHgx/7PdngGAtfc6QZFTsBpQfJaPV0JpTt6EFw1EafBchTCKARSJAgvS11SIIVyH2KJmtnCToR2nCERCDfz1OQtoXjP+KkoSpSioYA7B+uuSNSCZzqtbX5rHR20O7+RsQ3l6Km7Tiwu42ytMcnqYTs2SCNx6OUsoQwiQV3cRk/BiJxTwQtY6WsQceC2AoyU6Ft820+/jk7tZZ39EbQyLIrpBPMtMEghKaeFSYiF9j/F4/ERiY1qGx4AdDqZjqVTQhhtR+4QCTvswtXcC3H6+S83B194WfUW35/TzeXXXm3mbxE8bQyuPLLc7GbF2/69U7FY3bk+08y8/Yvm4qs/ttB0lPt+ZFo83HOrOfqVb7C/9wnWd7/Y3LxygQ7GZV3kghBmz57F0ZGVdFpwsMQouYwRrEu9ksPucg7BgbtmZXcTXrp1I06LWxtYy0RKkH+LNeyJlk/ZS8ziGwBhCQl2/o2fs/XdC0FIl7eKWAaxg62TswjwIBLI9SEKngs730b6ShQgWiRFyizXHRULIvFa44ISqw8Ov+h+7LduE3LW/cxZMgyxKGHKehese3D9DSJD6kVgGNER1EyFpnNKfRIHS5I38Nr1kEDcSXd4kt2OJ/fHQBILGmVOHmB3WyJxXnjjVRiD+8Ncat763ndj3ZOJLMGENGgYknUR+EzQF9tI+A33gHnw6CKYtL2FwYPJntT94yUSW9EShxxNJSaqVEskIld+dKoIOP3SV7IeeuLAgagBNy9cz0Y9l994O502gYg++PyzEe/zH11jU5o38WA9y0Ytt3CtuAaXLJN4qlVyfrF7iHUTuw/w3Stfy2C1D5VNK8wyKxDvMH7JpjYgEHawBDxw3JJBbA/qGT5aBjhIUAfUi7iXgyQrIP745G52bSWOFYSkircOkilmNlhM5P6QDuT7+0ciTTQRL96+Ec6pNMr8C+de1K2BWI4+3t2FsxZuWpiMsGNMAcIVdaowmhCIYw4Q3+shdpHaZA5oEyve0kezeDy7LV7ZgtpFX1Kc7vqTR0/EIqXEDJdX4jIDb4hZCYY/+LfB6TUBAwcJxMDTI5Mwpxe/TnjV46xqxBiieRtJ4f7v186eYzzmehI3x9Ej+Br7xR9nKwjWv+NQ6bjPbcZH+MadplTl3Mg0jKDFE/Fl5+i+3nn6pK8eH5HURtuiXMMF7eDun5IEFWSMCb1xfoe++jVWyR3ERDoCIt5tbl+60Ny+MpMVg/ufO9kcePH5Zi+bQbpfiQHWZt56i11e3425Vi/buMWrmsGh/Q1hxdq1f39z+JVvNEe+/rUMvpUOC+7XTfCDucsXWQR0HQJBjUJkuT+FG4Kqt+uu4Uq7SA84dHG70NUdCURTxplnMCK8rh3r6PcihXlYj1KXooa5k4MTa8t3bwbZBIdB80zrPMvgLjwBIBgPUaJceeNdF5IAt8BPmClF2l8fsBogrGoP9dL9ZogxxjILnNYxOtybnQ0DWcRo4C62TnQ6Yeie92PAZXwv68+VgLTXn+OSWx98GCIJAiNV+ahTC93mx9hA9sjXv4HKtA/mAZEQENyYzh+9925zm23CrbMu7+bnEt59z3+p2ffMSQiE0EqJIomEo91z1z5KH0VCVsKoZ3K5r+25BRYtnPL6Cf55fERSG1Ebjo5bzLYQipKETlbtGgLYx7/9bzTP/p2/jZq1pzn309ebmXfONB/+8PsgFhNyqACqSN/8R/8OptOT4E5f8/N/8b9l56wFEHwU7rQIF1wRwemYan+3+D1YlZ77e7/ffPUf/xMk0ySER4DumevN29/9E4hjJrv+ooiDCMXMG0JBp84aEXB8EEIYZL5DaSKnXO9lYEsHHz6mCRfkQi1xHCA3lag0DrigyGWqttGjID3qBls9m0YpZzhUI9G7b30G78bikmPL1u87QIwWfhVm9ex4w6gswzCXEVRPx3DC1PHb5p17iUx/a+ZKTNkbbjdhnchriHHdqb/3HTZYAtEhKAlOGG9A7GFM586Di0oiAlhANMIzih9nvQZ2HznaPP+d7xAkm28hyh4MJqvsGXKX8eASG+Dcw73HsupY8tnf/Z3my3//j5h0dK8S6sY7QxTduHCNsREqK671HkVy1sYX5vAgNOrbJ33+/ImkEoUt4bqjKvgckV2liQA2gsYkHXXjnfdRnd5XD2iuv/ceiHwlXMiB7q6pE83zf/v3QKiR5uaFi83Vd043Z37wg/gerdApcn7dsbUiFa9V5gAY8B964cXmlX/3nyJFULPo3Otvv9vcunChufr223j50vkMSl21WEKbUleuRWCl2CAD0l372MbA+LbUWRVpAfVsBFVijPkFA94xH93pK+s2MLTc3GZwm6jrvHG8M77/cIjEwfvt8+9EwixuGZcLZOPdwCAz3agbEid/Sh06uXKh9JCPC0cZS8tcJAYl8ABE7GZJ7lnpgFjL2qW/fjUMZO7Sxag9vahzfVr5gI8RGg2htP+55yNlJXBN4ctsy+eyWkMqaUHcMZVD8GE8jp1QXZUcvNcYsop5WwJT0iI7ILjJqMqz771KP5T2qGbaN84PrUEcEnBWZKK+aYEE4sCCxWlE13z40eIP8H+ax+dPJB9rjVyh/YVOSoc7uNQFRES/y8ByHt3ZnaruzFwEqEQqAQnHJk5gV2efeH53Zq5lT/hrDOJv49ZguB5NwJp4RQAwLIjm4HAXasTxr32jOfwiuz4xmFycvRl14BbzILcvnFP/SSdRkx0pIpHA6ewPJ9zMx5tsOYe5WSuPuwpLWA5qJZxy+AFXZgVCBdnDHKT5YoJ1c1QtcrZJd/E+HPv6mGDTUJB8QKoQitmQr/8kDDOt0rcylzqmU4r0MD5yj3ONBetEW1xevdvcuXwZFYuVl0g+iT7iTgIkT8ci7vRUnSxj3kUfdD2J8YwllKie1iF1Ah6ebSv18XvNv1tGmulluQBB5gwXtLlcVE1V0dSTtI5BlRqL7E8zd+0aKvV+8pF4BCuSCiLxvWtLPHbgmdv7/wQW9z96knePl0hsnH3t4WXbWF2r5UhjbAn8y+//ACsR6gGu6kCqk0ZT7D4Cuu07eSqD419+7/9uZs+fixpxj52WVF1ESDtOxHKSS114F3G1jr38SvN3/4N/xg5Uo81H5y807/34x83Zn/44OyDp5LjEeMdDBMuSUfLJYN884a4b82yFzCIkxzXWyQ5cZ4Wei6lceTjBtmI84qe50+W+WIjIN7PHaS8vIQiXB+uu4ez2c6g4gQUfqtpkPxGYwhZuGiGuFinJNXXr/BFmUVVbNZUxRDg4UrhncJTx2j2IYqFZwJBxg1jEC3Oz1JOVguHaToSWyVAZgOMCXW+s9xoqmC4C+mbdvUqABvas17VdyanatUQkGZlPIRL7DsuWIRrZ9np9Wbd5UKcXacw6/XtXcOthjLFMpExddIwxrBlYv7izr/6cZb1N8+If/lHgQ9G8M04w48qZq+xVf745ePJop7lfxIvHRyTd1N8hFOxbcFfVn0VE8Lk338Z6dT2IaieKRXJRueUQYngZl5IrDLDf+/6fwY0ul5ly3skB5ZBxvgOpRaxwIvBrZWm1ufjW283/8V/8l1wvgtwuRyXU6Y3ZIHPMuuHoIJ0l8k3WaStF2s7lA4Jiw/XtUZGWhKoYSyBRHZj7jWqF3xhNxEDTh15+mX1JWN+upYlnbtiZpgsLYnxFWlCXPgPBQVQpV4Ru25OxEXl2DtUtvhVBq8FD+I1inTPaiWrbAHMzCzevZ2fjRfZc1EBQF0vxWeAkrGRQUydOJcSpVrpeTNCuRly4znbNb59mfEC8MSSKKpF+axobJGYltURV5rhQz2iTaqib+JS6If2vsD4ea9WdmUv2XkmvmgaM5jC+7Gb2fZiJQ11X+iByYT6FL9c6xDhLGfZd6T9efAGPx0YkQY6HNNjOWgA495iXcH/AjAl4pmriXuxaiezbcUyGSgt1/LvsMbGCDpzZXBFHxA0CewnXQv3JegjycWGTSKXbhUHoNMU6X+GOvUoGOazI4/LbdEybl4hgnk6CpdN8zs+/ZEv+XvNecyh6/CAqlBnZTvN3/foIOxC7/4d+XWl/VctA0ujjEiLlO7FXOD3X3ksYFVHIr3NYsDn5vyUY62m41uVFxhiMQfpHbsWcvTALs9F8LcHZFr4ly9Tbb0eZ/8hmo5S1qEUPIlH6LWABK2oirVMiAn+fZ4wH/AvzsnyRGwJCzb135VyYnWMa+814ZZrFnbQtschop4UDLwluEQ/rO+fPRo0lJx4bYrXEGY5KSsqSvrbcb9OEwL8+fVrnx0YkNtouziHAbDc/ue8t9OZlJMlyJqmYK2Dwpz47wSB0hRhNLlyaZq5kBcTWWuTEXNQrEGqLOYDu9eZ2cOHQPmdlHFJqEFv82uI8psm5zF/4Ps59QTqqodCyOiDTFjdxZBSB+cUcy9l4wKo56VQTw/6cse/DzWJkwp11XT/BpGCVFqZJAyEmpQx1DeFRdlQW7+s1dZYwOoTitTACeZJHW89k2QLRNzIY4Xf36gx+ZRC9rjl84+Sdy2TXM8cDPMwrefgVzYCAp0+eJCYWg27S3foAIwlty1wGc0wyFoncGfcSstXxAlLE+ZLAh7bTzl5Usy1UyLlf/ixtc1zkbPs8TGyZYHSZnUfFgsICS+ugs+gyc0RXf/YTpDDr23FPWYMYVddch6NUKwwKuNn+FlfSBq+/AMdjI5K0LQ22oSIB3A1OvrWy3Uyho/egMvQ8B8eFE8Xu3+racj2BtYrF5RZjjwU4lDb+eOCCMA625VYijBIjqpdP4d4u+NEG72BQ1SGuFnBG1S2JULwpyFiRkrOEwXyA9RNZ6mY0I0iyPbivKN0GUWvQyHeIvkXqID1cVxlTCSDI73va4DMlnGVGgnjOO2AhQnitJBFO6iCec0gZSsyuw0e00Zl4GYnvVYUcX60uIjFV25J/azzg2kH/5LGTzd5nnmXG/2Qzh1o0e+5DBvXsXNsSsuOsSdqpFcqwRzH9At9CINSTI+MRpNbtmUvMvXyE+fZMiNW+0gOBmR++ZwwHrI2eIpxd7572QFiLWAbf+dnPeVfqpiQf0usAfzfX/AiH8msZibgiw/Gv7ejAJY+e+J/Pn0hskFDy4DqNLNhZgNY+EyjZMqEiislZKah1xI2L1uBGfei9w5s4KkpIIL+r6hTxmdhTreG5SFeP4jfkqjjVNDgtHdRDWJQQFKZQxwqqQ87+ZvAL0lmPSBI6JisJSaO6pPPdEKbdXuZlXGknEdmW9B2ERcvSHiVGEN735mV9kqftUZqUdHnmO+/bc6l7ixA8zrsKu06j6oXqEZY8iYvD8ZgMYoNtt+NYSD1sJy98qybYTEEcu/YfwMKnWzzFkkbpk6rLsJBoqlWFOGAm5Kc1LwYLEk0yz5SxCciv9UxzsNs3NNtwf4shjfuxb6xj1EBKaI6fOnkqVkv7Jilor2vqXY0oLB2rOcteKqHEVeqUfuzAtzQx3yelMHuKx+dPJDbGRtXO9rr9BQh0coiD9yJJ59rP7FlUGkGyRme5x3sPoWdEfjmTKoERA/Wd0vPUgaAMWBZvh9++eAH9HBeRduCpqlACspXqZIdZPFD3HMelAi6m567f2xEFeXeq7aOC7AURulWC6Om0yfZYt9K+LiKp7a1EYbqHXQdb/T6lJb/Uxdv7JEt5rxXN9SgpFyyNAyXqTaSpRAoSWs+Anj/Tz5xqRrFmubTW8gsTKZIrBAOCKjHWkbyaY8t4AgufUpA6T+Jf5XyIlqi5+UsQIy6LMKsQiHUm3RZSWtXPPUms39SJl9mPhfEP8zYm0R1+8fZsc/P2+RhjeliAZYML8tMGCDN1s37+hF2+pA7tmQcFxrl48n8eD5G0jbK5BRiKTJChBX6AUIHCOaZTJQiI3gNns5cniabRw2x8HbCGS6b3fU2u/uSKnPOY7/efONlsH3ceo0XcJOM9/+ICTn2sk/XIQbSU0jHcicgOVuwkr9s80mm1ru27Sjw76er3pY18vJOHBSVLkbfkXSphkrYc06QRXtRDmNW6FHRZAZk1TXtoaVtlfGEE99TZtqLW7n32+WbyyNFYlPSjCqxIr9u7UqOadat/1TbMSO8DPi+wBtkHYECO7Q4giTSsSCRarjJntK6Kq4QEdn6E2re6eDduRbsPHmz2Hj/ZKdPq72KyuAd43+q9wHeqW1jZ9UYA1m5R5yas26rEwjhSue0D+8gf/5/28fiIpKtlFdHSYUprFxVx+FeeEusUHErkyw+EEWAPEkjt8IpQIn+9Nic7JUfnooUx+ZGzKTrvS5IWSfO4Xovg/K9Ect+Z53xYxxnJTmQxs/ZX3luSBGhxvvPEvad871V7+J2Hz203l8KlQ1CW7wGCDTOvNOjgm7SDYwS2AEYStgxAGvN7pYZuL46htlzDkRcQiRKD8VOcGMlTlUp11LOqrCZdt4obnZ4OgSgJ/FaE1vNh/3MvMqa5nH5xHNOBJbU1rSrdLhwZEy2G6haY2GyYA6mHqLcqmX3N8BM3Go01qM6kLSor7bSt9WcJwqb+hMFTOh4vkdDA2um2L9c8C5LZ9zzQpaEAgwdKE9+XHs89N7HR+8zvva9HJ119EPRqb0yWD8opNRHgPpZDtS8rMZTn1ME0/FInOqy7o6r1xSiP+Txpyzc1/1KEeYgcJbvkmQJoW56Vevio885Lfp3WkZEtDjysD0TiQNetsixLV3+PsowYAwgfuva+rhmPyZpvCreHSDJoJo1hfAhHihE7TAiQo8Zyh9TRWDF19FgsThKGdZMROcu++8AhVK+FGEWcq7G2pb4u5WWcB2Eab6uMeWx/+RX1D58xpFPqBOwtLxY1xn7Zd7ISRoW3OadsS3mDAP4VAAAFHUlEQVT6x+e/HVx3m+y5itScK1IXVwvRgR4KF/Sa90oT0qVruI4kqfnV772veebSb9sjl1339XmAXm46SM9t51okbDu1dK4vlQRtF+V9F2Gk60pHphtzWe6D1DXvUqQF1av7r3ee7rSptlO4iMH+GKtpvYvvls/h7ve9Q20pMAGGnWsf1ecUxLUbehrlxLFfWI7tQhqRW9JqxSr5kL6rzoGNj7QE+o31thwPbuq4LkT9CTDrHseZt+WGQUkYrWVOKRKVy0y5LupZC7uu+pSCn9zfxy5JAmw7S8C0CKAZk36hPwCAyOh7u8111LnmMdwundnem76ka4FTn+fW780vf3O986cAucC4C+BeWqe8sB5+4bk+49YOTx19BZF4Mn2+Kdfd5sn6LlkldT6oV59+bvOsSGpedUySskEqD8db2dAzCE31IJoKV79N2S1sKiyTp89CQBKe6QrM0l5h761pauVrfVJq+QNkuu64TBrh1cLCR9SvJGvhmJe+F34m9Jo/lUi4DsGUTEhDOuCeNNzl8N1TPB4vkdSG2ch0AI3nWbqHi20XHPGudmYAY7r89w8ElJvyLNn5vj3qd+W2RRBvahIL6xw7NynHOrVcr3RIfc/Z/9bLi/vStJn5bd6XdDtFlOed+894UcruKov2pI4iIO0vMClwK3XnfeCGRKiN95uu68CKNJFAHRiSiP9FMUzWNsvCPrnmpM/RwsDr1IH7fGod87D8qe/8rF6bUGIqBCVR+E1bpvl05W0uT/t4MkRiK214e64dGpcMn6XjPRck8FGIKh1SesUOv48o6jdJ/Ah/2k4InqROpV61erWe3R2V6+4iauJ67n73616bp23jLBLZZn8pSmSy4gaQQJLU8UYlgNKSAi+rcR+8kB41XarY5vtJ1f1Ymx9M2N329rrCzBr4qEgP6pH7ivztS4gi6WVCWhU9+KgQTk2bTMq7p/j3yRFJdyMLBEuHBzYAw873uWef5W/5E3Wifebz+zq/K90nPa9JOh1vZ+w8rFf3n2sa6/Tg8bBnD6b5rPddeVd1yyDSDk1oeXLN3Ex7Lbx2asj7kqSk84sWniWd70sCT53vapquspPBJ/x5aDLVqfaF+XZg/cB1sky6tnSv+ZneXwb37TOffxGOp0MkteUPAUKAWzutTdfV7wUH/O6BNCbtdEz3u4eU0Q38zje1TvVcv6vn+vxJnNv21brZ/lzzvEgDK9ESehUnPMk78aoLYMnKB3lWCMp0Qb9uOJmlx6/a3vvwl5uu77pYEJLhvoQpIqW36Wsb/f4+AknKL8afp0sk3TDoAnI3wE3yMTA/rHNrXg971513MvxYjvXrL8651lmE9pqBrsgNvy7EYE0faOuDrQrRkCzPWyLxG/G2vjOb7qODtHmokluOnaty35JZ5217cV9n3Z9XJ0VpT9u+j6Wp7a7nnc+e2tUXh0geBQSfBkDfdSPPp6V9lDKfRlrr3tWeIBRt60asT0J2qxviqPUWJl1wue9dTfOQ8yelq8+tSyWk1PUheTz4qFP/h/XNw549mMETvv/NJJK/CUhfQED/TVX+1Pef0p4Own1qBrxsCeTTiOpvyuLB97XsSjB5311Xy6z3lUDrfXdmD3vW/f4pX///k0ieMlAfa/G/JkLdh9CfVtFPKufTkN38Pum77ufd159Why/Iu98SyRekIx57NSpi1vNnLfBRvn+UtJ+1Pk/guxgWn0A5vy3itxD4jYXAb4nkN7brflvxJwWB/w+BVh8aPT3LkgAAAABJRU5ErkJggg=="
                }
            ]
        }
    ],
    "temperature": 0.6,
    "stream": false
}
```

**Response:**

```json
{
    "id": "jobId-QmcJohc71DrHnVbXbLRt7drDyPSU1dyNTskU3a1yRR7zBu-jobState-ResultsSubmitted",
    "object": "chat.completion",
    "created": 1744411025,
    "model": "llava:7b",
    "system_fingerprint": "",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": " The image shows a creature that appears to be a blend of a frog and some form of robotic or mechanical structure. It has the body shape of a frog, with prominent eyes and limbs. The creature is depicted with a sleek, technologically advanced design, featuring metallic parts and futuristic elements. The color scheme includes blues, blacks, and greens, giving it a high-tech aesthetic. This creature could be an example of concept art for a video game or science fiction story. "
            },
            "finish_reason": "stop"
        }
    ],
    "usage": {
        "prompt_tokens": 600,
        "completion_tokens": 106,
        "total_tokens": 706
    }
}
```

### Embeddings

Use the embeddings endpoint to compute embeddings for user queries supported by the `nomic-embed-text` model. This endpoint is OpenAI compliant which means you can use it with the OpenAI SDK (see the end of the Embeddings section for a code example)

**Endpoint**

`POST /api/v1/embeddings`

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Request Parameters**

<table><thead><tr><th width="117">Parameter</th><th width="365.47265625">Description</th><th width="264.9296875">Type</th></tr></thead><tbody><tr><td><code>model</code><mark style="color:red;">*</mark></td><td>Model ID used to generate the response (e.g. <code>nomic-embed-text</code>). <strong>Required</strong>.</td><td><code>string</code></td></tr><tr><td><code>input</code><mark style="color:red;">*</mark></td><td>The input to create embeddings from. This can be either a single string or an array of strings. <strong>Required</strong></td><td><code>string or array of strings</code></td></tr></tbody></table>

**Request Sample (single input)**

```json
{
    "model": "nomic-embed-text",
    "input": "why is the sky blue?",
}
```

**Response Sample (single input)**

```json
{
    "object": "list",
    "data": [
        {
            "object": "embedding",
            "embedding": [
                0.009716383,
                0.0443843,
                -0.14053911,
                0.0011783413,
                0.031978477,
                0.1073299,
                -0.008574652,
                ...,
                -0.009498251,
                -0.041506674,
                0.020256031
            ],
            "index": 0
        },
    ],
    "model": "nomic-embed-text",
    "usage": {
        "prompt_tokens": 6,
        "total_tokens": 6
    }
}
```

**Request Sample (multiple input)**

```json
{
    "model": "nomic-embed-text",
    "input": ["why is the sky blue?", "why is the grass green?"],
}
```

**Response Sample (multiple input)**

```json
{
    "object": "list",
    "data": [
        {
            "object": "embedding",
            "embedding": [
                0.009716383,
                0.0443843,
                -0.14053911,
                0.0011783413,
                0.031978477,
                0.1073299,
                -0.008574652,
                ...,
                -0.009498251,
                -0.041506674,
                0.020256031
            ],
            "index": 0
        },
        {
            "object": "embedding",
            "embedding": [
               0.028126348,
                0.043248065,
                -0.18586768,
                0.03491587,
                0.055507593,
                0.12088179,
                -0.009062298
                ...,
                -0.035023507,
                -0.07451658,
                0.011851714
            ],
            "index": 1
        }

    ],
    "model": "nomic-embed-text",
    "usage": {
        "prompt_tokens": 12,
        "total_tokens": 12
    }
}
```

**Response Codes**

* `200 OK`: Request successful, stream begins
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `404 Not Found`: Requested model not found
* `500 Internal Server Error`: Server error processing request

**Example Code using the OpenAI SDK**

```typescript
import OpenAI from 'openai'

const openai = new OpenAI({
  baseURL: 'https://anura-testnet.lilypad.tech/api/v1/',
  apiKey: 'YOUR_KEY_HERE',
})

const embedding = await openai.embeddings.create({
  model: "nomic-embed-text",
  input: ["why is the sky blue?", "why is the grass green?"],
})
```

### Image Generation

The Anura API enables you to run stable diffusion jobs to generate images executed through our decentralized compute network. It's really easy to get started generating your own generative AI art using Anura through the endpoints we provide.

**Retrieve the list supported image generation models**

`GET /api/v1/image/models`&#x20;

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Request Parameters**

<table><thead><tr><th width="117">Parameter</th><th width="515.5">Description</th><th width="105.5">Type</th></tr></thead><tbody><tr><td><code>model</code><mark style="color:red;">*</mark></td><td>Model ID used to generate the response (e.g. <code>sdxl-turbo</code>). <strong>Required</strong>.</td><td><code>string</code></td></tr><tr><td><code>prompt</code><mark style="color:red;">*</mark></td><td>The prompt input to generate your image from (max limit of 1000 characters)</td><td><code>string</code></td></tr></tbody></table>

**Request Sample**

```bash
curl -X GET "https://anura-testnet.lilypad.tech/api/v1/image/models" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_api_key_here"
```

**Response**

```json
{
    "data": {
        "models": [
            "sdxl-turbo"
        ]
    },
    "message": "Retrieved models successfully",
    "status": 200
}
```

**Response Codes**

* `200 OK`: Request successful, stream begins
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `404 Not Found`: Requested model not found
* `500 Internal Server Error`: Server error processing request

Currently we support `sdxl-turbo`; however, we are always adding new models, so stay tuned!

**Generate an AI Image**

`POST /api/v1/image/generate`

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Request Parameters**

<table><thead><tr><th width="117">Parameter</th><th width="515.5">Description</th><th width="105.5">Type</th></tr></thead><tbody><tr><td><code>model</code><mark style="color:red;">*</mark></td><td>Model ID used to generate the response (e.g. <code>sdxl-turbo</code>). <strong>Required</strong>.</td><td><code>string</code></td></tr><tr><td><code>prompt</code><mark style="color:red;">*</mark></td><td>The prompt input to generate your image from (max limit of 1000 characters)</td><td><code>string</code></td></tr></tbody></table>

**Request Sample**

```json
{
    "prompt": "A spaceship parked on a lilypad",
    "model": "sdxl-turbo"
}
```

Alternatively you can also make the same request through a curl command and have the image be output to a file on your machine

```bash
curl -X POST https://anura-testnet.lilypad.tech/api/v1/image/generate \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your_api_key_here" \
  -d '{"prompt": "A spaceship parked on a lilypad", "model": "sdxl-turbo"}' \
  --output spaceship.png
```

The result of running this command will be the creation of the `spaceship.png` file in the directory you ran the command from.&#x20;

**Response**

This endpoint will return the raw bytes value of the image that was generated which you can output to a file (like shown in the curl command above) or place it in a buffer to write to a file in your app, e.g.&#x20;

```javascript
const fs = require("fs");
const fetch = require("node-fetch");

async function generateImage() {
  const response = await fetch("https://anura-testnet.lilypad.tech/api/v1/image/generate", {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      "Authorization": "Bearer your_api_key_here"
    },
    body: JSON.stringify({
      prompt: "A spaceship parked on a lilypad",
      model: "sdxl-turbo"
    }),
  });
  
  if (!response.ok) {
    console.error(`Error generating image: StatusCode: ${response.status} Error: ${response.message}`);
    return;
  }

  const buffer = await response.buffer();
  fs.writeFileSync("spaceship.png", buffer);
}

generateImage();
```

**Note:** Should you ever need to know what the corresponding Job Offer ID for image generation, it is provided in the response header as `Job-Offer-Id`

**Response Codes**

* `200 OK`: Request successful, stream begins
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `404 Not Found`: Requested model not found
* `500 Internal Server Error`: Server error processing request

### Video Generation

The Anura API enables you to run long running jobs to generate videos executed through our decentralized compute network. It's really easy to get started generating your own videos using Anura through the endpoints we provide.&#x20;

**Note**: Video generation can take anywhere between 4-8 mins to produce a video

**Retrieve the list supported video generation models**

`GET /api/v1/video/models`&#x20;

Currently we support `wan2.1`; however, we are always adding new models, so stay tuned!

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Request Sample**

```bash
curl -X GET "https://anura-testnet.lilypad.tech/api/v1/video/models" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_api_key_here"
```

**Response**

```json
{
    "data": {
        "models": [
            "wan2.1"
        ]
    },
    "message": "Retrieved models successfully",
    "status": 200
}
```

**Response Codes**

* `200 OK`: Request successful
* `401 Unauthorized`: Invalid or missing API key
* `500 Internal Server Error`: Server error processing request

**Send out a request to create an AI generated video**

`POST /api/v1/video/create-job`

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Request Parameters**

<table><thead><tr><th width="166.12890625">Parameter</th><th width="486.68359375">Description</th><th width="105.5">Type</th></tr></thead><tbody><tr><td><code>model</code><mark style="color:red;">*</mark></td><td>Model used to generate the response (e.g. <code>wan2.1</code>). <strong>Required</strong>.</td><td><code>string</code></td></tr><tr><td><code>prompt</code><mark style="color:red;">*</mark></td><td>The prompt input to generate your video from (max limit of 1000 characters). <strong>Required</strong>.</td><td><code>string</code></td></tr><tr><td><code>negative_prompt</code></td><td>An optional field to specify to the model what to <strong>exclude</strong> from the generated scene</td><td><code>string</code></td></tr></tbody></table>

**Request Sample**

```json
{
    "prompt": "Two frogs sit on a lilypad, animatedly discussing the wonders and quirks of AI agents. As they ponder whether these digital beings can truly understand their froggy lives, the serene pond serves as a backdrop to their lively conversation.",
    "negative_prompt": "Dull colors, grainy texture, washed-out details, static frames, incorrect lighting, unnatural shadows, distorted faces, artifacts, low-resolution elements, flickering, blurry motion, repetitive patterns, unrealistic reflections, overly simplistic backgrounds, three legged people, walking backwards.",
    "model": "wan2.1"
}
```

**Response**

This endpoint will return an `job_offer_id` which is an unique identifier corresponding to the job that's running to create your video. What you'll want to do with this id is pass it into our `/video/results` endpoint (see below) which will provide you the output as a `webp` file or report that the job is still running. In the latter case, you then can continue to call the endpoint at a later time to eventually retrieve your video. As mentioned in the beginning of this section, video generation can take anywhere between 4-8 mins to complete.

```json
{
    "status": 200,
    "message": "Video job created successfully",
    "data": {
        "job_offer_id": "<your-job-offer-id-here>"
    }
}
```

**Response Codes**

* `200 OK`: Request successful, stream begins
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `404 Not Found`: Requested model not found
* `500 Internal Server Error`: Server error processing request

**Retrieve your video**

`GET /api/v1/video/results/:job_offer_id`&#x20;

<table><thead><tr><th width="166.12890625">Parameter</th><th width="434.75390625">Description</th><th width="105.5">Type</th></tr></thead><tbody><tr><td><code>job_offer_id</code><mark style="color:red;">*</mark></td><td>The id returned to you in the video creation request i.e <code>/api/v1/video/create-job</code><strong>Required</strong>.</td><td><code>string</code></td></tr></tbody></table>

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Response**

If the video is still in the process of being generated you will see a response that looks like the following:

```json
{
    "status": 102,
    "message": "Request is still processing",
    "data": {
        "job_offer_id": "<job-offer-id>",
        "job_state": "DealAgreed"
    }
}
```

**Response Codes**

* `102 Processing`: Request is still processing the creation of the video
* `200 OK`: Request successful
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `500 Internal Server Error`: Server error processing request

However, once the video has be generated you'll be returned the video in `webp` format with its raw bytes which you can save to a file in the following manner:

```bash
curl -X GET "https://anura-testnet.lilypad.tech/api/v1/video/results/<your-job-offer-id>" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_api_key_here" \
--output video.webp
```

The result of the above command will be the `video.webp` file being saved in the directory from which you ran it from:

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F4CM3vgenZitV5mVjCFVJ%2Ffrogs_chatting_listening.webp?alt=media&amp;token=9eed6853-c320-44e8-8a69-20a3f3117ff8" alt=""><figcaption><p>Two frogs sit on a lilypad, animatedly discussing the wonders and quirks of AI agents. As they ponder whether these digital beings can truly understand their froggy lives, the serene pond serves as a backdrop to their lively conversation.</p></figcaption></figure>

### Audio Generation

The Anura API enables you to generate audio from text executed through our decentralized compute network. It's really easy to get started generating your own audio using Anura through the endpoints we provide.&#x20;

**Note**: Audio generation can take anywhere between 40 seconds to 3 mins to complete depending on the input length

**Retrieve the list supported audio generation models**

`GET /api/v1/audio/models`&#x20;

Currently we support `kokoro`; however, we are always adding new models, so stay tuned!

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Request Sample**

```bash
curl -X GET "https://anura-testnet.lilypad.tech/api/v1/audio/models" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_api_key_here"
```

**Response**

```json
{
    "data": {
        "models": [
            "kokoro"
        ]
    },
    "message": "Retrieved models successfully",
    "status": 200
}
```

**Response Codes**

* `200 OK`: Request successful
* `401 Unauthorized`: Invalid or missing API key
* `500 Internal Server Error`: Server error processing request

**Send out a request to create an AI generated audio**

`POST /api/v1/audio/create-job`

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Request Parameters**

<table><thead><tr><th width="166.12890625">Parameter</th><th width="486.68359375">Description</th><th width="105.5">Type</th></tr></thead><tbody><tr><td><code>model</code><mark style="color:red;">*</mark></td><td>Model used to generate the response (e.g. <code>kokoro</code>). <strong>Required</strong>.</td><td><code>string</code></td></tr><tr><td><code>input</code><mark style="color:red;">*</mark></td><td>The prompt input to generate your audio from (max limit of 420 characters). <strong>Required</strong>.</td><td><code>string</code></td></tr><tr><td><code>voice</code><mark style="color:red;">*</mark></td><td>The voice to use when generating the audio sample. Possible values are <code>heart</code>, <code>puck</code>, <code>fenrir</code>, and <code>bella</code><strong>Required</strong>.</td><td><code>string</code></td></tr></tbody></table>

**Voice samples**

**Heart**

{% file src="/files/rvsPCGCclwGQUSacFmR8" %}
Heart Voice Sample
{% endfile %}

**Puck**

{% file src="/files/naekS3sbxWOjkAfhe43w" %}
Puck Voice Sample
{% endfile %}

**Fenrir**

{% file src="/files/PM6d0qTKa1pStoUT8Lly" %}
Fenrir Voice Sample
{% endfile %}

**Bella**

{% file src="/files/cwBRb56hkuFm5GlYeArL" %}
Bella Voice Sample
{% endfile %}

**Request Sample**

```json
{
    "input": "Hello my name is Heart and this is AI speech generated from text using the Kokoro module running on the Lilypad Network",
    "voice": "heart",
    "model": "kokoro"
}
```

**Response**

This endpoint will return an `job_offer_id` which is an unique identifier corresponding to the job that's running to create your audio. What you'll want to do with this id is pass it into our `/audio/results` endpoint (see below) which will provide you the output as a `wav` file or report that the job is still running. In the latter case, you then can continue to call the endpoint at a later time to eventually retrieve your audio. As mentioned in the beginning of this section, audio generation can take anywhere between 40 seconds to 3 mins to complete.

```json
{
    "status": 200,
    "message": "Audio job created successfully",
    "data": {
        "job_offer_id": "QmTmTWxffQrosK2nb3a4oeeZd9KRMUGwApEUFhioFUe4Y9"
    }
}
```

**Response Codes**

* `200 OK`: Request successful
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `404 Not Found`: Requested model not found
* `500 Internal Server Error`: Server error processing request

**Retrieve your video**

`GET /api/v1/audio/results/:job_offer_id`&#x20;

<table><thead><tr><th width="166.12890625">Parameter</th><th width="434.75390625">Description</th><th width="105.5">Type</th></tr></thead><tbody><tr><td><code>job_offer_id</code><mark style="color:red;">*</mark></td><td>The id returned to you in the audio creation request i.e <code>/api/v1/audio/create-job</code><strong>Required</strong>.</td><td><code>string</code></td></tr></tbody></table>

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Response**

If the audio is still in the process of being generated you will see a response that looks like the following:

```json
{
    "status": 102,
    "message": "Request is still processing",
    "data": {
        "job_offer_id": "QmTmTWxffQrosK2nb3a4oeeZd9KRMUGwApEUFhioFUe4Y9",
        "job_state": "DealAgreed"
    }
}
```

**Response Codes**

* `102 Processing`: Request is still processing the creation of the audio
* `200 OK`: Request successful
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `500 Internal Server Error`: Server error processing request

However, once the audio has be generated you'll be returned the audio in `wav` format with its raw bytes which you can save to a file in the following manner:

```bash
curl -X GET "https://anura-testnet.lilypad.tech/api/v1/audio/results/<your-job-offer-id>" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_api_key_here" \
--output audio.wav
```

### Web Search

The Anura API provides developers with a web search capability enabling you to add a powerful tool to your AI Agent building arsenal. LLM's are only as great as their training data and are taken to the next level when provided with additional context from the web. With web search you can power your AI Agent workflow with live web search data providing your LLM the most up to date information on the latest on goings in the world.

It's easy to get started searching the web through the Anura API using our endpoint:

`POST /api/v1/websearch`

**Request Headers**

* `Content-Type: application/json`<mark style="color:red;">\*</mark>
* `Authorization: Bearer YOUR_API_KEY`<mark style="color:red;">\*</mark>

**Request Parameters**

<table><thead><tr><th width="193.38671875">Parameter</th><th width="515.5">Description</th><th width="91.625">Type</th></tr></thead><tbody><tr><td><code>query</code><mark style="color:red;">*</mark></td><td>The web search query you wish to execute</td><td><code>string</code></td></tr><tr><td><code>number_of_results</code><mark style="color:red;">*</mark></td><td>The number of search results you want returned (limited to 1 to 10 inclusive)</td><td><code>number</code></td></tr></tbody></table>

**Request Sample**

```json
{
    "query": "What's the Lilypad Network?",
    "number_of_results" : 3
}
```

**Response Sample**

The response will include the following fields:

<table><thead><tr><th width="176.0234375">Field</th><th width="557.90234375">Description</th></tr></thead><tbody><tr><td><code>results</code></td><td>The array of search results where each result object is made up of the strings: <code>title</code>, <code>url</code> and <code>description</code></td></tr><tr><td><code>related_queries</code></td><td>An array of strings containing similar queries based on the one you supplied</td></tr><tr><td><code>count</code></td><td>The number of search results returned</td></tr></tbody></table>

```json
{
    "results": [
        {
            "title": "Lilypad Network",
            "url": "https://lilypad.tech",
            "description": "Lilypad Network Lilypad offers a seamless and efficient way to access the computing power you need for AI and other demanding tasks—no need to invest in expensive hardware or navigate complex cloud setups. Simply submit your job; our decentralized network connects you with the best available resources. Benefit from competitive pricing, secure ..."
        },
        {
            "title": "Lilypad Network - internet-scale off-chain distributed compute solution",
            "url": "https://blog.lilypadnetwork.org",
            "description": "Verifiable, truly internet-scale distributed compute network Efficient off-chain computation for AI & ML DataDAO computing The next frontier of web3. Follow. ... Check out the docs https://docs.lilypad.tech/lilypad! Lilypad Builder-verse! Devlin Rocha. 4 min read. Fuel the Future by Building on Lilypad and Accelerate Open Source AI. Alex Mirran."
        },
        {
            "title": "What is the Lilypad Decentralized Compute Network?",
            "url": "https://blog.lilypadnetwork.org/what-is-the-lilypad-decentralized-compute-network",
            "description": "Lilypad democratizes AI high-performance computing, offering affordable, scalable solutions for researchers and startups. Follow. Follow. What is the Lilypad Decentralized Compute Network? A Crowdsourced Network for HPC Tasks. Lindsay Walker"
        }
    ],
    "related_queries": [
        "Lilypad Tech",
        "LilyPad github",
        "Lilypad website",
        "Lilypad AI",
        "Lily pad Minecraft server",
        "Lilypad crypto",
        "LilyPad Arduino"
    ],
    "count": 3
}
```

**Response Codes**

* `200 OK`: Request successful, stream begins
* `400 Bad Request`: Invalid request parameters
* `401 Unauthorized`: Invalid or missing API key
* `404 Not Found`: Requested model not found
* `500 Internal Server Error`: Server error processing request

### Jobs

* `GET /api/v1/jobs/:id` - Get status and details of a specific job

#### Get Status/Details of a Job

You can use another terminal to check job status while the job is running.

```bash
curl -X GET "https://anura-testnet.lilypad.tech/api/v1/jobs/{job_id}" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_api_key_here"
```

#### Get Outputs from a Job

Once your job has run, you should get output like this:

```
data: {
    "id": "cmpl-e654be2df70700d27c155d4d",
    "object": "text_completion",
    "created": 1738614839,
    "model": "llama2",
    "choices": [{
        "text": "<output text here>
        "finish_reason": "stop"
    }]
}
```

### Cowsay

* `POST /api/v1/cowsay` - Create a new cowsay job
  * Request body: `{"message": "text to display"}`
* `GET /api/v1/cowsay/:id/results` - Get results of a cowsay job


# AI Model Marketplace

Introduction to running AI models on Lilypad with the Module Marketplace

Lilypad makes it easy to deploy and run AI models with an AI model hub called the [Lilypad Module Marketplace](https://github.com/Lilypad-Tech/awesome-Lilypad?tab=readme-ov-file#modules).&#x20;

Currently setup in the [awesome-lilypad github repo](https://github.com/Lilypad-Tech/awesome-Lilypad?tab=readme-ov-file#modules), the Module Marketplace makes it easy for AI model creators to distribute their work and for the community to quickly get started running a wide range of open source AI models.

There are two ways that participants in the network can interact with Lilypad Modules: As *Job Creators* who run modules, or as *Module Creators* who create modules.

## Running Modules

Running Jobs (participating as a 'Job Creator') is just a few simple steps:

* Use the Lilypad [AI inference API](/lilypad/developer-resources/inference-api) or [install our CLI](/lilypad/quickstart/cli/installation)
* Choose a module to run and send a job request

### **Example Modules**

Try out some of the Lilypad team modules:

* [Cowsay](/lilypad/developer-resources/module-marketplace/build-a-job-module/cowsay)
* [Llama2](/lilypad/developer-resources/module-marketplace/build-a-job-module/llama2)
* [Stable Diffusion Turbo Pipeline](/lilypad/developer-resources/module-marketplace/build-a-job-module/stable-diffusion-turbo-pipeline)

Additional modules from the community are available in the [awesome-lilypad repo](https://github.com/Lilypad-Tech/awesome-Lilypad?tab=readme-ov-file#modules).

## Creating Modules

A [Lilypad Module](https://docs.lilypad.tech/lilypad/developer-resources/module-marketplace/build-a-job-module) is a standard containerized (Docker) process for running compute workloads on Lilypad.&#x20;

Modules can be developed by any member of the community (participating as a 'Module Creator'). If there is an AI model (or other compute job) you wish to run using the Lilypad Network resources, create a new Module and [add it to the Module Marketplace](https://blog.lilypadnetwork.org/lilypad-module-creator-rewards-beta#heading-submitting-a-module)!

### Developer Resources

* [Build a Job Module](/lilypad/developer-resources/module-marketplace/build-a-job-module)
* [create-lilypad-module](/lilypad/developer-resources/module-marketplace/create-lilypad-module)
* [Module Builder Guide](https://blog.lilypadnetwork.org/lilypad-module-builder-guide)


# Build a Job Module

How to build your own compute job for Lilypad

A Lilypad module is a Git repository that allows you to perform various tasks using predefined templates and inputs. This guide will walk you through creating a Lilypad module, including defining a JSON template, handling inputs, and following best practices.

For a more in-depth look at building modules, refer to this [end-to-end guide](https://blog.lilypadnetwork.org/lilypad-module-builder-guide).

{% hint style="info" %}
If you're new to Docker, consider exploring this [step-by-step tutorial](https://docs.docker.com/get-started/) on creating, building, and running a Docker image for a simple Hello World style application.
{% endhint %}

## Modules on Lilypad

Below are a few examples of modules you can run on Lilypad. From language models to image generators and fun utilities, the network supports a growing list of AI modules.

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>Llama2</td><td>An open-source language model, used for generating and understanding text.</td><td><a href="/lilypad/developer-resources/module-marketplace/build-a-job-module/llama2">Llama2</a></td></tr><tr><td>Stable Diffusion Turbo Pipeline</td><td>A fast, high-performance version of Stable Diffusion XL for generating high-quality images with minimal latency</td><td><a href="/lilypad/developer-resources/module-marketplace/build-a-job-module/stable-diffusion-turbo-pipeline">Stable Diffusion Turbo Pipeline</a></td></tr><tr><td>Cowsay</td><td>A playful CLI tool that displays text as if spoken by an ASCII-art cow</td><td><a href="/lilypad/developer-resources/module-marketplace/build-a-job-module/cowsay">Cowsay</a></td></tr></tbody></table>

To view the full list of available modules on Lilypad, please check out [the awesome-lilypad repo](https://github.com/Lilypad-Tech/awesome-Lilypad?tab=readme-ov-file#modules)!

## Module Structure

Start by creating a Git repository for your Lilypad module. The module's versions will be represented as Git tags. Below is the basic structure of a Lilypad Module.

```
your-module/
├── model-directory            # Stores locally downloaded model files 
├── download_model.[py/js/etc] # Script to download model files locally
├── requirements.txt           # Module dependencies
├── Dockerfile                 # Container definition
├── run_script.[py/js/etc]     # Main execution script
├── lilypad_module.json.tmpl   # Lilypad configuration
└── README.md                  # Documentation
```

## Prepare Your Model

* Download model files
* Handle all dependencies (`requirements.txt`)
* Implement input/output through environment variables
* Write outputs to `/outputs` directory

### **1. Download the model locally**

To use a model offline, you first need to download it and store it in a local directory. This guarantees that your code can load the model without requiring an internet connection. Here's a simple process to achieve this:

1. Install required libraries
2. Use a script to download the model (eg: `python download_model.py`)
3. Verify that the model files are in your directory

{% hint style="info" %}
The method to download and save a model and tokenizer may vary based on the model's architecture and the framework you are using. Always refer to the documentation of the specific model to ensure compatibility and proper usage.
{% endhint %}

```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

def download_model():
    model_name = "<namespace>/<model_identifier>"
    # Ensure you have a directory named 'model' in your current working directory or specify a path
    tokenizer = AutoTokenizer.from_pretrained(model_name)
    model = AutoModelForSeq2SeqLM.from_pretrained(model_name)

    # Save the tokenizer and model
    tokenizer.save_pretrained('./model')
    model.save_pretrained('./model')

if __name__ == "__main__":
    download_model()
```

### **2. Create Run Script (run\_model.py for example) that will be used in conjunction with Docker**

```python
import os
import json
from transformers import AutoModel, AutoTokenizer

def main():
    # Load model and tokenizer from local directory
    model_path = '/model'  # Path to the local model directory
    tokenizer = AutoTokenizer.from_pretrained(model_path)
    model = AutoModel.from_pretrained(model_path)

    # Get inputs from environment variables
    input_var = os.environ.get('INPUT_VAR', 'default')
    
    # Your model code here
    result = your_model_function(input_var, model, tokenizer)
    
    # Save outputs
    output_path = '/outputs/result.json'
    with open(output_path, 'w') as f:
        json.dump({'result': result}, f)

if __name__ == "__main__":
    main()

```

### 3. Create a Dockerfile that functions with your run script

```dockerfile
# Use specific base image
FROM base-image:version

# Set working directory
WORKDIR /workspace

# Install dependencies
RUN apt-get update && apt-get install -y \
    your-dependencies && \
    rm -rf /var/lib/apt/lists/*

# Install model requirements
RUN pip install your-requirements

# Environment variables for running offline and using the local model 
# HF_HOME points to the directory where the model code is
ENV HF_HOME=/model
ENV TRANSFORMERS_OFFLINE=1

# Create necessary directories
RUN mkdir -p /outputs

# Copy execution script
COPY run_script.* /workspace/

# Set entrypoint
ENTRYPOINT ["command", "/workspace/run_script"]
```

### 4. Build and Publish Image

To make sure your Docker image is compatible with Lilypad, you need to define the architecture explicitly during the build process. This is particularly important if you are building the image on a system like macOS, which uses a different architecture (`darwin/arm64`) than Lilypad's infrastructure (`linux/amd64`).

The examples below are for building, tagging and pushing an image to DockerHub, but you can use any platform you prefer for hosting the image.&#x20;

For Linux: `docker buildx build -t <USERNAME>/<MODULE_NAME>:<MODULE_TAG> --push .`

For MacOS:&#x20;

<pre class="language-bash"><code class="lang-bash">docker buildx build \
<strong>--platform linux/amd64 \
</strong>-t &#x3C;USERNAME>/&#x3C;MODULE_NAME>:&#x3C;MODULE_TAG> \
--push \
.
</code></pre>

### 5. Create a lilypad\_module.json.tmpl Template

```json
{
    "machine": {
        "gpu": 1,          # Set to 0 if GPU not needed
        "cpu": 1000,       # CPU allocation
        "ram": 8000        # Minimum RAM needed to run the module
    },
    "gpus": [ { "vram": 24576 }, { "vram": 40960 } ] # VRAM in MBs. Solver will default to largest one 
    "job": {
        "APIVersion": "V1beta1",
        "Spec": {
            "Deal": {
                "Concurrency": 1
            },
            "Docker": {
                "Entrypoint": ["command", "/workspace/run_script"],
                "WorkingDirectory": "/workspace",
                "EnvironmentVariables": [
                    # Environment variables with defaults
                    {{ if .var_name }}"VAR_NAME={{ js .var_name }}"{{ else }}"VAR_NAME=default_value"{{ end }}
                ],
                # Specify the Docker image to use for this module
                "Image": "repo-owner/repo-name:tag"
            },
            "Engine": "Docker",
            "Network": {
                "Type": "None"
            },
            "Outputs": [
                {
                    "Name": "outputs",
                    "Path": "/outputs"
                }
            ],
            "PublisherSpec": {
                "Type": "ipfs"
            },
            "Resources": {
                "GPU": "1"    # Must match machine.gpu
            },
            "Timeout": 1800
        }
    }
}
```

## Environment Variables

Format in template:&#x20;

```
{{ if .variable }}"VARNAME={{ js .variable }}"{{ else }}"VARNAME=default"{{ end }}
```

Usage in CLI:

```
lilypad run repo:tag -i variable=value
```

## Formatting your module run command

During development, you will need to use the Git hash to test your module. This allows you to verify that your module functions correctly and produces the expected results.

Below is a working Lilypad module run cmd for reference. (you can use this to run a Lilypad job within the Lilypad CLI):

### Test Module before running on Lilypad

Use the following command syntax to run your Module on Lilypad Testnet.

```
lilypad run github.com/Lilypad-Tech/module-sdxl:6cf06f4038f1cff01a06c4eabc8135fd9835a78a --web3-private-key <your-private-key> -i prompt="a lilypad floating on a pond"
```

If the job run appears to be stuck after a few minutes (sometimes it takes time for the Module to download to the RP node), cancel the job and try again. Open a ticket in [Discord](https://discord.com/channels/1212897693450641498/1230231823674642513) with any issues that persist.&#x20;

{% hint style="info" %}
If many jobs have been run on the machine previosuly, clear `Lilypad` from the `/tmp` folder locally and try running the job again.
{% endhint %}

### Run Module on Lilypad

```bash
lilypad run github.com/noryev/module-sdxl-ipfs:ae17e969cadab1c53d7cabab1927bb403f02fd2a -i prompt="your prompt here"
```

## Examples

Here are some example Lilypad modules for reference:

* [**Cowsay**](https://github.com/Lilypad-Tech/lilypad-module-cowsay)**:** Lilypad "Hello World" example
* [**Llama2**](/lilypad/developer-resources/module-marketplace/build-a-job-module/llama2): Text to text
* [**SDXL-turbo pipeline**](/lilypad/developer-resources/module-marketplace/build-a-job-module/stable-diffusion-turbo-pipeline): Text to image generation
* [**create-lilypad-module**](/lilypad/developer-resources/module-marketplace/create-lilypad-module)&#x20;

Deprecated examples:

* [**lora-training**](https://github.com/Lilypad-Tech/lilypad-module-lora-training)**:** An example module for LoRa training tasks.
* [**lora-inference**](https://github.com/Lilypad-Tech/lilypad-module-lora-inference)**:** An example module for LoRa inference tasks.
* [**duckdb**](https://github.com/Lilypad-Tech/lilypad-module-duckdb)**:** An example module related to DuckDB.

These examples can help you understand how to structure your Lilypad modules and follow best practices.

## Conclusion

In this guide, we've covered the essential steps to create a Lilypad module, including defining a JSON template, handling inputs, and testing your module. By following these best practices, you can build reliable and reusable modules for Lilypad.&#x20;

For more information and additional examples, refer to the official Lilypad documentation and the Cowsay example module.


# Llama2

Run Llama 2 on the Lilypad network

These instructions provide steps for running the Llama2 module on the Lilypad network using Docker and the Lilypad CLI. Find the module repo [here](https://github.com/noryev/module-llama2/tree/main).

## Getting Started <a href="#getting-started" id="getting-started"></a>

Before running `llama2`, make sure you have the [Lilypad CLI installed](/lilypad/quickstart/cli/installation) on your machine and your private key environment variable is set. This is necessary for operations within the Lilypad network.

```bash
export WEB3_PRIVATE_KEY=<YOUR_PRIVATE_KEY>
```

Learn more about installing the Lilypad CLI and running a Lilypad job with this [video guide](https://www.youtube.com/watch?v=RBECCMl_fco).

## Run Llama2

#### Run Llama2 <a href="#run-sdxl-turbo" id="run-sdxl-turbo"></a>

```bash
lilypad run github.com/noryev/module-llama2:6d4fd8c07b5f64907bd22624603c2dd54165c215 -i prompt="your prompt here"
```

Example:

```bash
lilypad run github.com/noryev/module-llama2:6d4fd8c07b5f64907bd22624603c2dd54165c215 -i prompt="what is a giant sand trout on arrakis?"
```

#### Notes <a href="#notes" id="notes"></a>

* Ensure you have the necessary permissions and resources to run Docker containers with GPU support.
* The module version (6d4fd8c07b5f64907bd22624603c2dd54165c215) may be updated. Check for the latest version before running.
* Adjust port mappings and volume mounts as needed for your specific setup.

### Output <a href="#sdxl-output" id="sdxl-output"></a>

To view the results in a local directory, navigate to the local folder provided by the job result.

```
open /tmp/lilypad/data/downloaded-files/<fileID>
```

<br>


# Stable Diffusion Turbo Pipeline

A Lightweight Stable Diffusion Module for Lilypad

These instructions provide steps for running the SDXL Turbo Pipeline module on the Lilypad network using Docker and Lilypad CLI. Find the module repo [here](https://github.com/noryev/module-sdxl-ipfs).

## Getting Started

Before running `sdxl turbo`, make sure you have the [Lilypad CLI installed](/lilypad/quickstart/cli/installation) on your machine and your private key environment variable is set. This is necessary for operations within the Lilypad network.

```bash
export WEB3_PRIVATE_KEY=<YOUR_PRIVATE_KEY>
```

Learn more about installing the Lilypad CLI and running a Lilypad job with this [video guide](https://www.youtube.com/watch?v=RBECCMl_fco).

## Run SDXL Turbo

```
lilypad run github.com/Lilypad-Tech/module-sdxl:d6a89ed92f4e798459b2990340669da00c56c80c -i prompt="your prompt here"
```

Example:

```
lilypad run github.com/Lilypad-Tech/module-sdxl:d6a89ed92f4e798459b2990340669da00c56c80c -i prompt="a spaceship parked on a mountain"
```

### Notes

* Ensure you have the necessary permissions and resources to run Docker containers with GPU support.
* The module version (`ae17e969cadab1c53d7cabab1927bb403f02fd2a`) may be updated. Check for the latest version before running.
* Adjust port mappings and volume mounts as needed for your specific setup.

### Output

To view the results in a local directory, navigate to the local folder.

```
open /tmp/lilypad/data/downloaded-files/<fileID>
```


# Cowsay

Run a cowsay job

`cowsay` is a simple, text-based program originally written for Unix-like operating systems that generates ASCII pictures of a cow with a speech bubble containing a specified message.&#x20;

This module was created as a "Hello World" for the Lilypad Network!

## Getting Started

Before running `cowsay`, make sure you have the [Lilypad CLI installed](/lilypad/quickstart/cli) on your machine and your private key environment variable is set. This is necessary for operations within the Lilypad network.

```bash
export WEB3_PRIVATE_KEY=<YOUR_PRIVATE_KEY>
```

## Run cowsay

Once you've installed the CLI, run the `cowsay` command:

```bash
lilypad run cowsay:v0.0.4 -i Message="hello, lilypad"  
```

### Output

```bash
⠀⠀⠀⠀⠀⠀⣀⣤⣤⢠⣤⣀⠀⠀⠀⠀⠀
⠀⠀⠀⠀⢴⣿⣿⣿⣿⢸⣿⡟⠀⠀⠀⠀⠀    ██╗     ██╗██╗  ██╗   ██╗██████╗  █████╗ ██████╗
⠀⠀⣰⣿⣦⡙⢿⣿⣿⢸⡿⠀⠀⠀⠀⢀⠀    ██║     ██║██║  ╚██╗ ██╔╝██╔══██╗██╔══██╗██╔══██╗
⠀⢰⣿⣿⣿⣿⣦⡙⣿⢸⠁⢀⣠⣴⣾⣿⡆    ██║     ██║██║   ╚████╔╝ ██████╔╝███████║██║  ██║
⠀⣛⣛⣛⣛⣛⣛⣛⠈⠀⣚⣛⣛⣛⣛⣛⣛    ██║     ██║██║    ╚██╔╝  ██╔═══╝ ██╔══██║██║  ██║
⠀⢹⣿⣿⣿⣿⠟⣡⣿⢸⣮⡻⣿⣿⣿⣿⡏    ███████╗██║███████╗██║   ██║     ██║  ██║██████╔╝
⠀⠀⢻⣿⡟⣩⣾⣿⣿⢸⣿⣿⣌⠻⣿⡟⠀    ╚══════╝╚═╝╚══════╝╚═╝   ╚═╝     ╚═╝  ╚═╝╚═════╝ v2.13.0
⠀⠀⠀⠉⢾⣿⣿⣿⣿⢸⣿⣿⣿⡷⠈⠀⠀
⠀⠀⠀⠀⠀⠈⠙⠛⠛⠘⠛⠋⠁⠀ ⠀⠀⠀   Decentralized Compute Network  https://lilypad.tech

🌟  Lilypad submitting job
2025-03-05T12:56:38-06:00 WRN ../runner/work/lilypad/lilypad/cmd/lilypad/utils.go:63 > failed to get GPU info: gpuFillInfo not implemented on darwin
2025-03-05T12:56:38-06:00 INF ../runner/work/lilypad/lilypad/pkg/web3/sdk.go:209 > Connected to arbitrum-sepolia-rpc.publicnode.com
2025-03-05T12:56:38-06:00 INF ../runner/work/lilypad/lilypad/pkg/jobcreator/run.go:27 > Public Address: 0xB86bCAe21AC95BCe7a49C057dC8d911033f8CB7c
Enumerating objects: 42, done.
Counting objects: 100% (22/22), done.
Compressing objects: 100% (4/4), done.
Total 42 (delta 18), reused 19 (delta 18), pack-reused 20 (from 1)
💌  Deal agreed. Running job...
🤔  Results submitted. Awaiting verification...
🤔  Results submitted. Awaiting verification...
✅  Results accepted. Downloading result...
🆔  Data ID: QmP2SQttNC3Hrh2xpY7bNHzV2jHq7MbfLahRC46DVzn5rG

🍂 Lilypad job completed, try 👇
    open /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q
    cat /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q/stdout
    cat /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q/stderr
```

To view the results in a local directory, navigate to the local folder.

```bash
open /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q
```

Here, you can view the `stdout` and `stderr` as well as the `outputs` folder for the run:

```bash
~ % cat /tmp/lilypad/data/downloaded-files/QmQHrsiAuzTLn5VU6jg5LoXBRrAkEVRKiYeJE29w54gg9Q/stdout
 ________________ 
< hello, lilypad >
 ---------------- 
        \   ^__^
         \  (oo)\_______
            (__)\       )\/\
                ||----w |
                ||     ||
```


# create-lilypad-module

Use create-lilypad-module to create Lilypad modules

`create-lilypad-module` is an officially supported package that provides a simple scaffolding for building Lilypad modules. It offers a modern Docker setup with minimal configuration.

{% hint style="info" %}
A Lilypad module is a Git repository that allows you to perform various tasks using predefined templates and inputs.
{% endhint %}

{% hint style="warning" %}
The following guide is using the "Hugging Face" template, which is more advanced.

If you are new to module building, it is highly recommended you use the "Ollama" template first.

There is currently not a guide in our docs for the "Ollama" template. We will work on adding one soon, but the directions in the README should be sufficient.
{% endhint %}

## Overview

* [Getting Started](/lilypad/developer-resources/module-marketplace/create-lilypad-module/getting-started): install and run `create-lilypad-module`
* [Folder Structure](/lilypad/developer-resources/module-marketplace/create-lilypad-module/folder-structure): output and explanation of `create-lilypad-module` files
* [Configuration](/lilypad/developer-resources/module-marketplace/create-lilypad-module/configuration): requirements and explanations of Lilypad module configuration
* [Creating Your Module](/lilypad/developer-resources/module-marketplace/create-lilypad-module/creating-your-module): a step-by-step guide on how to create a simple Lilypad module using `create-lilypad-module`


# Getting Started

Getting started with \`create-lilypad-module\`

## Installation

First, you'll need to install the package:

```sh
pip install create-lilypad-module
```

If you've previously installed `create-lilypad-module`, you should to ensure that you're using the latest version:

```sh
pip install --upgrade create-lilypad-module
```

Now run `create-lilypad-module`:

```sh
create-lilypad-module
```

The CLI will ask for the name of your project. Alternatively, you can run:

```sh
create-lilypad-module project_name
cd project_name
```

Output

```
project_name
├── config
│   └── constants.py
├── scripts
│   ├── docker_build.py
│   ├── download_models.py
│   └── run_module.py
├── src
│   └── run_inference.py
├── .dockerignore
├── .env
├── .gitignore
├── Dockerfile
├── lilypad_module.json.tmpl
├── README.md
└── requirements.txt
```


# Folder Structure

The folder structure output from using \`create-lilypad-module\`

## Output

After creation, your project should look like this:

```
project_name
├── config
│   └── constants.py
├── scripts
│   ├── docker_build.py
│   ├── download_models.py
│   └── run_module.py
├── src
│   └── run_inference.py
├── .dockerignore
├── .env
├── .gitignore
├── Dockerfile
├── lilypad_module.json.tmpl
├── README.md
└── requirements.txt
```

For the module to run, **these files must exist with exact filenames**:

* `src/run_inference.py`
  * The `Dockerfile` `ENTRYPOINT`.
  * If you change this files name or location, you must also update the `ENTRYPOINT` in your `Dockerfile` and `lilypad_module.json.tmpl` file to match.
* `config/constants.py`
  * The configuration file that stores the `DOCKER_REPO`, `DOCKER_TAG`, `MODULE_REPO`, and `TARGET_COMMIT`.
  * If you change this files name or location, you must also update the `import` statements in `scripts/docker_build.py` and `scripts/run_module.py`.
* `Dockerfile`
  * Required to build your module into a Docker image, and push the image to Docker Hub where it can be accessed by Lilypad Network.
* `requirements.txt`
  * Used by the `Dockerfile` to install dependencies required by your module.
  * Technically, this file can be deleted or renamed, but this naming convention is highly recommended as an industry standard best practice.
* `lilypad_module.json.tmpl`
  * The Lilypad configuration file.

You can delete or rename the other files.

You may create subdirectories inside `src`. For faster builds and smaller Docker images, only files inside `src` are copied by Docker. You need to put any files required to run your module inside `src`, otherwise Docker won’t copy them.

You can create more top-level directories. They will not be included in the final Docker image so you can use them for things like documentation.

If you have Git installed and your project is not part of a larger repository, then a new repository will be initialized resulting in an additional top-level `.git` directory.


# Configuration

Configure your Lilypad module

## Configure Your Module

After bootstrapping your module, additional configuration is required to run it.

### `.env`

```
WEB3_PRIVATE_KEY = ""
```

#### `WEB3_PRIVATE_KEY`

> 🚨 **DO NOT SHARE THIS KEY** 🚨

The private key for the wallet that will be used to run the job.

This is required to run the module on Lilypad Network.

A new development wallet is highly recommended to use for development. The wallet must have enough LP tokens and Arbitrum Sepolia ETH to fund the job.

* [Funding your wallet](/lilypad/quickstart/cli/setting-up-your-wallet)

### `config/constants.py`

```python
DOCKER_REPO = ""
MODULE_REPO = ""
TARGET_COMMIT = ""
```

#### `DOCKER_REPO`

The Docker Hub repository storing the container image of the module code.

This is required to push the image to Docker Hub and run the module on Lilypad Network.

e.g. `"<dockerhub_username>/<dockerhub_image>"`

#### `DOCKER_TAG`

The specific tag of the `DOCKER_REPO` containing the module code.

Default: `"latest"`

#### `MODULE_REPO`

The URL for the GitHub repository storing the `lilypad_module.json.tmpl` file. The visibility of the repository must be public.

The `lilypad_module.json.tmpl` file points to a `DOCKER_REPO` and Lilypad runs the module from the image.

e.g. `"github.com/<github_username>/<github_repo>"`

#### `TARGET_COMMIT`

The git branch or commit hash that contains the `lilypad_module.json.tmpl` file you want to run.

Use `git log` to easily find commit hashes.

Default: `"main"`

### Available Scripts

Your module will be bootstrapped with some handy scripts to help you download the model(s) for your module, build and push Docker images, and run your module locally or on Lilypad Network. Some additional configuration may be required.

In the project directory, you can run:

#### `python -m scripts.download_models`

A basic outline for downloading a model from [Hugging Face](https://huggingface.co/) is provided, but the structure of the script and the methods for downloading a model can differ between models and libraries. It’s important to tailor the process to the specific requirements of the model you're working with.

Most (but not all) models that utilize machine learning use the [🤗 Transformers](https://huggingface.co/docs/transformers/index) library, which provides APIs and tools to easily download and train pretrained models.

* [Learn more about downloading models from Hugging Face](https://huggingface.co/docs/hub/en/models-downloading)
* [Learn more about the 🤗 Transformers library](https://huggingface.co/docs/hub/en/transformers)

No matter which model you are using, be sure to thoroughly read the documentation to learn how to properly download and use the model locally.

#### `python -m scripts.docker_build`

Builds and optionally publishes a Docker image for the module to use.

For most use cases, this script should be sufficient and won't require any configuration or modification (aside from setting your `DOCKER_REPO` and `DOCKER_TAG`).

{% hint style="info" %}
In the modules `Dockerfile`, you'll find 3 COPY instructions.

```dockerfile
COPY requirements.txt .
COPY src /src
COPY models /models
```

These instructions copy the `requirements.txt` file, the `src` directory, and the `models` directory from your local machine into the Docker image. It's important to remember that any modifications to these files or directories will necessitate a rebuild of the module's Docker image to ensure the changes are reflected in the container.
{% endhint %}

**`--push` Flag**

Running the script with `--push` passed in pushes the Docker image to Docker Hub.

**`--no-cache` Flag**

Running the script with `--no-cache` passed in builds the Docker image without using the cache. This flag is useful if you need a fresh build to debug caching issues, force system or dependency updates, pull the latest base image, or ensure clean builds in CI/CD pipelines.

#### `python -m scripts.run_module`

This script is provided for convenience to speed up development. It is equivalent to running the Lilypad module with the provided input and private key (unless running the module locally, then no private key is required). Depending on how your module works, you may need to change the default behavior of this script.

**`--local` Flag**

Running the script with `--local` passed in runs the Lilypad module Docker image locally instead of on Lilypad's Network.

**`--demonet` Flag**

Running the script with `--demonet` passed in runs the Lilypad module Docker image on Lilypad's Demonet.

### `lilypad_module.json.tmpl`

The default `lilypad_module.json.tmpl` file is below. Make sure to update the Docker Image to point to your Docker Hub image with the correct tag.

> The default `lilypad_module.json.tmpl` should work for low complexity modules. If your module requires additional resources (such as a GPU) make sure to configure the applicable fields.

```json
{
  "machine": { "gpu": 1, "cpu": 8000, "ram": 16000 },
  "gpus": [{ "vram": 24576 }]
  "job": {
    "APIVersion": "V1beta1",
    "Spec": {
      "Deal": { "Concurrency": 1 },
      "Docker": {
        "Entrypoint": [
          "/app/src/run_model", {{ .request }}
        ],
        "Image": "DOCKER_HUB_USERNAME/DOCKER_IMAGE@INDEX_DIGEST"
      },
      "Engine": "Docker",
      "Network": { "Type": "None" },
      "Outputs": [{ "Name": "outputs", "Path": "/outputs" }],
      "Resources": { "GPU": "1", "CPU": "8", "Memory": "16Gb" },
      "Timeout": 1800,
      "Verifier": "Noop"
    }
  }
}
```

#### Template Fields

* **Machine:** Specifies the system resources.
* GPUs: Specifies the minimum VRAM required.
* **Job:** Specifies the job details.
  * **APIVersion:** Specifies the API version for the job.
  * **Metadata:** Specifies the metadata for the job.
  * **Spec:** Contains the detailed job specifications.
    * **Deal:** Sets the concurrency to 1, ensuring only one job instance runs at a time.
    * **Docker:** Configures the Docker container for the job
      * **WorkingDirectory:** Defines the working directory of the Docker image.
      * **Entrypoint:** Defines the command(s) to be executed in the container as part of its initial startup runtime.
      * **EnvironmentVariables:** This can be utilised to set env vars for the containers runtime, in the example above we use Go templating to set the `INPUT` variable dynamically from the CLI.
      * **Image:** Specifies the image to be used (`DOCKERHUB_USERNAME`/`IMAGE`:`TAG`).
    * **Engine:** Sets the container runtime (Default: `"Docker"`).
    * **Network:** Specifies that the container does not require networking (Default: `"Type": "None"`).
    * **Outputs:** Specifies name and path of the directory that will store module outputs.
    * **Resources:** Specify additional resources.
    * **Timeout:** Sets the maximum duration for the job. (Default: `600` \[10 minutes]).


# Creating Your Module

Create your Lilypad module

## Getting Started

This guide will walk you through creating a basic sentiment analysis module using [`create-lilypad-module`](https://pypi.org/project/create-lilypad-module/) and [`distilbert/distilbert-base-uncased-finetuned-sst-2-english`](https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english) (which will be referred to as Distilbert from now on). We will be referring back to the Hugging Face page throughout this guide, so it's best to keep it open and accessible.

Input:

```sh
lilypad run --network demonet github.com/DevlinRocha/lilypad-module-sentiment:main --web3-private-key 0ec38dd1ee0898dae8460b269859b4fb3cb519b35d82014c909ec4741c790831 -i input="LILYPAD IS AWESOME"
```

Output:

```json
{
  "input": "LILYPAD IS AWESOME",
  "sentiment": "POSITIVE",
  "status": "success"
}
```

* [View the final source code for the module](https://github.com/DevlinRocha/lilypad-module-sentiment)

If you prefer to follow along with a video guide, you can view our live workshop below! 👇

{% embed url="<https://youtu.be/A_JYvV7k_JA?t=142>" fullWidth="false" %}
create-lilypad-module live workshop
{% endembed %}

### Prerequisites

To build and run a module on Lilypad Network, you'll need to have the [Lilypad CLI](/lilypad/quickstart/cli), [Python](https://www.python.org/) and [Docker](https://www.docker.com/) on your machine, as well as [GitHub](https://github.com/) and [Docker Hub](https://hub.docker.com/) accounts.

For this guide, we'll be using [`create-lilypad-module`](https://pypi.org/project/create-lilypad-module/) which requires [`pip`](https://pip.pypa.io/) and uses [`Python`](https://www.python.org/).

## Downloading The Model

The first thing you'll need for your module is a local model to use.

A basic outline for downloading a model from [Hugging Face](https://huggingface.co/) is provided in `scripts/download_models.py`. The structure of the script and the methods for downloading a model can differ between models and libraries. It’s important to tailor the process to the specific requirements of the model you're working with.

You can get started by attempting to run the `download_models.py` script.

```sh
python -m scripts.download_models
```

Since the script hasn't been properly configured yet, it will return an error and point you to the file.

```sh
❌ Error: Model download script is not configured.
👉 /scripts/download_models.py
```

Open `scripts/download_models.py` and you will see some `TODO` comments with instructions. Let's go through them each in order. You can remove each `TODO` comment after completing the task.

```python
# TODO: Update ../requirements.txt
# from transformers import AutoTokenizer, AutoModelForSequenceClassification
```

First we have a reminder to update our `requirements.txt` file, which is used by the `Dockerfile` to install the module's dependencies. In the next line is a commented out `import` statement.

To find the dependencies that our model requires, we can refer back to Distilbert's Hugging Face page and click on the "Use this model" dropdown, where you will see the [🤗 Transformers](https://huggingface.co/docs/hub/en/transformers) library as an option. Click it.

{% hint style="info" %}
Most (but not all) models that utilize machine learning use the 🤗 Transformers library, which provides APIs and tools to easily download and train pretrained models.

* [Learn more about downloading models from Hugging Face](https://huggingface.co/docs/hub/en/models-downloading)
* [Read the `🤗 Transformers` documentation](https://huggingface.co/docs/transformers/index)
  {% endhint %}

You should see a handy modal explaining how to use the model with the `Transformers` library. For most models, you'd want to use this. However, Distilbert has a specific tokenizer and model class. Close the modal and scroll to the [How to Get Started With the Model](https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english#how-to-get-started-with-the-model) section of the model card. We're going to use this instead.

For now, let's look at the top 2 lines of the provided code block:

```python
import torch
from transformers import DistilBertTokenizer, DistilBertForSequenceClassification
```

Notice that `torch` is also being used. Copy the `transformers` import statement and paste it over the existing import statement in our `download_models.py` file.

Now open `requirements.txt`:

```txt
# torch==2.4.1
# transformers==4.47.1
```

These are 2 of the most common libraries when working with models. Similar to the `import` statement in the `download_models.py` file, they are provided by default for convenience, but commented out because although they are common, not every model will use them.

Since this model happens to use both of these libraries, we can uncomment both lines and close the file after saving.

{% hint style="info" %}
`torch` is a collection of APIs for extending [PyTorch](https://pytorch.org/)’s core library of operators.

* [Learn more about using `datasets` with `Pytorch`](https://huggingface.co/docs/datasets/en/use_with_pytorch)
* [Read the `torch` documentation](https://pytorch.org/docs/stable/index.html)
  {% endhint %}

Return to the `download_models.py` file, and look for the next `TODO` comment.

```python
# TODO: Set this to your model's Hugging Face identifier
MODEL_IDENTIFIER = ""
```

If we take a look at the [Distilbert Hugging Face page](https://huggingface.co/distilbert/distilbert-base-uncased-finetuned-sst-2-english), we can use the copy button next to the name of the module to get the `MODULE_IDENTIFIER`. Paste that in as the value.

For our use case, it should look like this:

```python
MODEL_IDENTIFIER = "distilbert/distilbert-base-uncased-finetuned-sst-2-english"
```

You're almost ready to download the model. All you need to do now is replace the following 2 lines after the `TODO` comment:

```python
# TODO: Initialize `model` and `tokenizer`
# tokenizer = AutoTokenizer.from_pretrained(MODEL_IDENTIFIER)
# model = AutoModelForSequenceClassification.from_pretrained(MODEL_IDENTIFIER)
```

Instead of using `AutoTokenizer` and `AutoModelForSequenceClassification`, replace those with the `DistilBertTokenizer` and `DistilBertForSequenceClassification` we imported.

```python
tokenizer = DistilBertTokenizer.from_pretrained(MODEL_IDENTIFIER)
model = DistilBertForSequenceClassification.from_pretrained(MODEL_IDENTIFIER)
```

The script is now configured! Try running the command again.

```sh
python -m scripts.download_models
```

The `models` directory should now appear in your project. 🎉

{% hint style="info" %}
No matter which model you are using, be sure to thoroughly read the model's documentation to learn how to properly download and use the model locally.
{% endhint %}

## Building Your Model

Now for the fun part, it's time to start using the model!

This time we'll get started by running the `run_module` script.

```sh
python -m scripts.run_module
```

You should see an error with some instructions.

```sh
❌ Error: No job configured. Implement the module's job before running the module.
        1. Implement job module
                👉 /src/run_inference.py
        2. Delete this code block
                👉 /scripts/run_module.py
```

#### 1. Implement Job Module

Let's tackle the `run_inference.py` script first. This is where your modules primary logic and functionality should live. There is a `TODO` comment near the top of the file.

```python
# TODO: Update ../requirements.txt
# import torch
# from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
```

We've already updated the `requirements.txt` file, so we can skip that step. Go ahead and uncomment the `import` statements and replace the `transformers` line with the `DistilBertTokenizer` and `DistilBertForSequenceClassification`.

We should refer back to the "How to Get Started With the Model" section of Distilbert's model card to figure out how to use the model.

```python
tokenizer = DistilBertTokenizer.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")
model = DistilBertForSequenceClassification.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")

inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
with torch.no_grad():
    logits = model(**inputs).logits

predicted_class_id = logits.argmax().item()
model.config.id2label[predicted_class_id]
```

Let's implement this into our `run_inference` script. Scroll down to the `main()` function and you'll see another `TODO` comment.

```python
# TODO: Initialize `model` and `tokenizer`
# tokenizer = AutoTokenizer.from_pretrained(MODEL_DIRECTORY)
# model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_DIRECTORY)
```

Same as before, uncomment and replace `AutoTokenizer` with `DistilBertTokenizer` and `AutoModelForSeq2SeqLM` with `DistilBertForSequenceClassification`. This is now functionally identical to the first 2 lines of code from Distilbert's example.

Below that, the `tokenizer` and `model` are passed into the `run_job()` function. Let's scroll back up and take a look at the function. This is where we'll want to implement the rest of the code from Distilbert's example. The `inputs` are already functionally identical, so let's adjust the `output`.

From the Distilbert model card, copy all of the code below the `inputs` variable declaration, and paste it over the `output` variable declaration in your modules code.

```python
inputs = tokenizer(
    input,
    return_tensors="pt",
    truncation=True,
    padding=True,
)

with torch.no_grad():
    logits = model(**inputs).logits

predicted_class_id = logits.argmax().item()
model.config.id2label[predicted_class_id]

return output
```

All we need to do from here is set the `output` to the last line we pasted.

```python
output = model.config.id2label[predicted_class_id]

return output
```

That's everything we'll need for the modules source code!

#### 2. Delete Code Block

We still need to finish step 2 that the error in the console gave us earlier. Open the `run_module.py` script.

Find the `TODO` comment and delete the code block underneath.

```python
# TODO: Remove the following print and sys.exit statements and create the module job.
print(
    "❌ Error: No job configured. Implement the module's job before running the module.",
    file=sys.stderr,
    flush=True,
)
print("\t1. Implement job module")
print("\t\t👉 /src/run_inference.py")
print("\t2. Delete this code block")
print("\t\t👉 /scripts/run_module.py")
sys.exit(1)
```

## Preparing Your Module

Before you are able to run your module, we need to build the Docker image. You can run the following command:

```sh
python -m scripts.docker_build
```

You should see the following response in the console:

```sh
❌ Error: DOCKER_REPO is not set in config/constants.py.
```

Open the `constants.py` file, it should look like this:

```python
# TODO: Set the Docker Hub repository before pushing the image.
# Example: "devlinrocha/lilypad-module-sentiment"
DOCKER_REPO = ""

# TODO: Set the tag for the Docker image.
# Example: "latest", "v1.0", or a commit SHA
DOCKER_TAG = "latest"

# TODO: Set the GitHub repository URL where your module is stored.
# Example: "github.com/devlinrocha/lilypad-module-sentiment".
MODULE_REPO = ""

# TODO: Specify the target branch name or commit hash.
# Example: "main" or "c3ed392c11060337cae010862b1af160cd805e67"
TARGET_COMMIT = "main"
```

For now, we'll be testing the module locally, so all we need to worry about is the `DOCKER_REPO` variable. We'll use `MODULE_REPO` when it's time to run the module on Lilypad Network. For help or more information, view the configuration documentation

You should be able to successfully build the Docker image now.

{% hint style="info" %}
In the modules Dockerfile, you'll find 3 COPY instructions.

```dockerfile
COPY requirements.txt .
COPY src /src
COPY models /models
```

These instructions bring the `requirements.txt` file, the `src` directory, and the `models` directory into the Docker image. It's important to remember that any modifications to these files or directories will necessitate a rebuild of the module's Docker image to ensure the changes are reflected in the container.
{% endhint %}

## Running Your Module

It's finally time to see your module in action.

#### Local

Let's start by running it locally.

```sh
python -m scripts.run_module --local
```

The CLI should ask you for an input. Enter whatever you like and hit enter. The module will analyze the sentiment of your input and output the results at `outputs/result.json`.

```json
{
  "input": "LILYPAD IS AWESOME",
  "result": "POSITIVE",
  "status": "success"
}
```

You just used a local LLM! 🎉

#### Lilypad Network

Before you can run the module on Lilypad Network, you'll need to push the Docker image to Docker Hub.

```sh
python -m scripts.docker_build --push
```

While the Docker image is being built and pushed, you should configure the rest of the variables in `constants.py`. Make sure that you push your code to a public GitHub repository.

{% hint style="info" %}
Since these variables are only used in scripts and not in any `src` code that gets used in the Docker image we won't need to rebuild after making these changes.
{% endhint %}

The last thing we'll need to do is edit the Lilypad module configuration file, `lilypad_module.json.tmpl`. For the purposes of this module, the default configuration is mostly correct. However, the `"Image"` field needs to be configured.

Replace the default value with your Docker Hub username, module image, and tag.

```json
"Image": "devlinrocha/lilypad-module-sentiment:latest"
```

Once your Docker image is pushed to Docker Hub and your most recent code is pushed to a public GitHub repository, you can test your module on Lilypad's DemoNet by replacing the `--local` flag with `--demonet`

```sh
python -m scripts.run_module --demonet
```

> You can also remove the `--demonet` flag and supply your `WEB3_PRIVATE_KEY` to run the module on Lilypad's IncentiveNet.

You just used an LLM on Lilypad's decentralized network! 🎉

Now anyone who has the Lilypad CLI installed can also run your module:

```sh
lilypad run github.com/<MODULE_REPO>:<TARGET_COMMIT> --web3-private-key <WEB3_PRIVATE_KEY>
```


# Stable Diffusion (SDXL0.9 & 1.0)

Run a Stable Diffusion Text to Image Job

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fm0rOscLk0ZAko168PwXC%2Fcli-sdxl.gif?alt=media&amp;token=911ca8e2-e228-4994-8bd8-61b72e8d6fa8" alt=""><figcaption><p>Run a SDXL job on the Lilypad Network</p></figcaption></figure>

## Overview

Generically, stable diffusion is what happens when you put a couple of drops of dye into a bucket of water. Given time, the dye randomly disperses and eventually settles into a uniform distribution which colours all the water evenly.

In computer science, you define rules for your (dye) particles to follow and the medium this takes place in.

Stable Diffusion is a machine learning model used for text-to-image processing (like Dall-E) and based on a diffusion probabilistic model that uses a transformer to generate images from text.

## Getting Started

### Prerequisites

Before running `sdxl`, make sure you have the [Lilypad CLI installed](https://docs.lilypad.tech/lilypad/lilypad-testnet/install-run-requirements) on your machine and your private key environment variable is set. This is necessary for operations within the Lilypad network.&#x20;

```bash
export WEB3_PRIVATE_KEY=<YOUR_PRIVATE_KEY>
```

Learn more about installing the Lilypad CLI and running SDXL with this [video guide](https://www.youtube.com/watch?v=RBECCMl_fco).

### Run SDXL v0.9 or 1.0

When running SDXL pipelines in Lilypad, you have the choice between using the Base model or the Refiner model. Each serves a unique purpose in the image generation process:

* **Base Model**: This is the primary model that generates the initial image based on your input prompt. It focuses on the broad aspects of the image, capturing the main theme and essential elements. The Base model is faster and uses less computational power.
* **Refiner Model**: This model takes the image from the Base model and enhances it. It refines details, improves textures, and adjusts colors to increase the visual appeal and realism of the image. The Refiner model is used when you need higher quality and more detailed images.

&#x20;To run SDXL Pipeline in Lilypad, you can use the following commands:

#### SDXL v0.9

Base:

```bash
lilypad run sdxl-pipeline:v0.9-base-lilypad3 -i Prompt="an astronaut floating against a white background"
```

Refiner:

```bash
lilypad run sdxl-pipeline:v0.9-refiner-lilypad3 -i Prompt="an astronaut floating against a white background"
```

#### SDXL 1.0

Base:

```bash
lilypad run sdxl-pipeline:v1.0-base-lilypad3 -i Prompt="an astronaut floating against a white background"
```

Refiner:

```bash
lilypad run sdxl-pipeline:v1.0-refiner-lilypad3 -i Prompt="an astronaut floating against a white background"
```

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FBLIUFIesaN67rE3dNfrQ%2Fcli-sdxl.gif?alt=media&amp;token=ef7d0e49-6ab8-4786-9b6b-53df8b22dc27" alt=""><figcaption></figcaption></figure>

### SDXL Output

<figure><img src="https://github.com/noryev/lilypad-docs/raw/main/lilypad/.gitbook/assets/sdxl_execution.png" alt=""><figcaption><p>Output from a Lilypad job</p></figcaption></figure>

To view the results in a local directory, navigate to the local folder.

```bash
open /tmp/lilypad/data/downloaded-files/QmZuE29GJVmenRUh72FQDgkMUT1Zdp967oEJvzjaDwGGVoResults of SDXL job on Output Directory
```

In the **/outputs** folder, you'll find the image:

To view the results on IPFS, navigate to the IPFS CID result output.&#x20;

{% hint style="warning" %}
Please be patient! IPFS can take some time to propagate and doesn't always work immediately.
{% endhint %}

```
https://ipfs.io/ipfs/QmVng1jkMxE9ep4k8mYiiCiWaCRiRLvCeo6bJRXirhz1dZ
```

<figure><img src="https://github.com/noryev/lilypad-docs/raw/main/lilypad/.gitbook/assets/sdxl_result_output.png" alt=""><figcaption><p>SDXL job output</p></figcaption></figure>

As Lilypad modules are currently deterministic, running this command with the same text prompt will produce the same image, since the same seed is also used (the default seed is 0).

## Tuning an output

To change an image output, pass in a different seed number:

{% hint style="info" %}
See this [beginner-friendly article](https://aituts.com/stable-diffusion-seed/) on how seeds work for more info on this
{% endhint %}

Lilypad can run SDXL v0.9 or SDXL v1.0 with the option to add tunables to improve or change the model output.&#x20;

If you wish to specify more than one tunable, such as the number of steps, simply add more `-i` flags. For example, to improve the quality of the image generated add "Steps=x" with x = (Min: 5. Max: 200):

{% hint style="info" %}
See the options and tunables section (below) for more information on what tunables are available.
{% endhint %}

```bash
`lilypad run sdxl-pipeline:v0.9-base-lilypad3 -i Prompt='a gigantic lilypad shaped space station' -i Steps=150` 
```

<figure><img src="https://github.com/noryev/lilypad-docs/raw/main/lilypad/.gitbook/assets/sdxl_result_output2.png" alt=""><figcaption><p>Output with a different seed</p></figcaption></figure>

### Options and tunables continued

The following tunables are available. All of them are optional, and have default settings that will be used if you do not provide them.

| Name        | Description                              | Default                           | Available options                                                                                                                                                                                                                                                                                         |
| ----------- | ---------------------------------------- | --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `Prompt`    | A text prompt for the model              | "question mark floating in space" | Any string                                                                                                                                                                                                                                                                                                |
| `Seed`      | A seed for the model                     | 42                                | Any valid non-negative integer                                                                                                                                                                                                                                                                            |
| `Steps`     | The number of steps to run the model for | 50                                | Any valid non-negative integer from 5 to 200 inclusive                                                                                                                                                                                                                                                    |
| `Scheduler` | The scheduler to use for the model       | `normal`                          | `normal`, `karras`, `exponential`, `sgm_uniform`, `simple`, `ddim_uniform`                                                                                                                                                                                                                                |
| `Sampler`   | The sampler to use for the model         | `euler_ancestral`                 | `"euler"`, `"euler_ancestral"`, `"heun"`, `"heunpp2"`, `"dpm_2"`, `"dpm_2_ancestral"`, `"lms"`, `"dpm_fast"`, `"dpm_adaptive"`, `"dpmpp_2s_ancestral"`, `"dpmpp_sde"`, `"dpmpp_sde_gpu"`, `"dpmpp_2m"`, `"dpmpp_2m_sde"`, `"dpmpp_2m_sde_gpu"`, `"dpmpp_3m_sde"`, `"dpmpp_3m_sde_gpu"`, `"ddpm"`, `"lcm"` |
| `Size`      | The output size requested in px          | `1024`                            | `512`, `768`, `1024`, `2048`                                                                                                                                                                                                                                                                              |
| `Batching`  | How many images to produce               | `1`                               | `1`, `2`, `4`, `8`                                                                                                                                                                                                                                                                                        |

See the usage sections for the runner of your choice for more information on how to set and use these variables.

Learn more about this Lilypad module on [Github](https://github.com/Lilypad-Tech/lilypad-module-sdxl-pipeline).

## Looking for Smart Contracts?

Check out [our smart contracts docs](/lilypad/archive/lilypad-smart-contracts)!


# Stable Diffusion Video (SDV1.0 & 1.1)

Run a Stable Diffusion Text to Video Job

## Overview

Based on ComfyUI, the SDV Pipeline modules for Lilypad allow you generate videos from text prompts on Lilypad using Stable Diffusion Video and related models.&#x20;

The SDV Pipeline modules are designed to take your text prompt, generate a still frame using SDXL, then use that as the input to the SDV model, producing an APNG (animated PNG), WebP video, and an MP4 video all in one go.

## Getting Started

### Prerequisites

Before running `sdv`, make sure you have the [Lilypad CLI installed](https://docs.lilypad.tech/lilypad/lilypad-milky-way-testnet/install-run-requirements) on your machine and your private key environment variable is set. This is necessary for operations within the Lilypad network.

```bash
export WEB3_PRIVATE_KEY=<YOUR_PRIVATE_KEY>
```

### Running SDV v1.0 or 1.1

To run SDV v1.0 or 1.1 Pipeline in Lilypad, you can use the following commands:

#### SDV 1.0

```bash
lilypad run sdv-pipeline:v1.0-lilypad2 -i ImageSeed="696721260153400" -i Prompt="an astronaut floating against a white background" -i Steps=200 -i VideoSteps=70
```

#### SDV 1.1

```bash
lilypad run sdv-pipeline:v1.1-lilypad2 -i ImageSeed="696721260153400" -i Prompt="an astronaut floating against a white background" -i Steps=200 -i VideoSteps=70
```

### SDV Output

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FqQ2l6kqOAqVAYO3oTrqm%2FsdvJob.png?alt=media&amp;token=57ed8108-fb45-453a-8bf8-49fdde19a193" alt=""><figcaption></figcaption></figure>

To view the results in a local directory, navigate to the local folder.

```bash
open /tmp/lilypad/data/downloaded-files/QmbjEKwZn5vSBoArudUEAPJDEiesVPk5tNkYxFvNkdFs6t
```

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FRb7xKte2tHygYaaRErUI%2FsdvOutput.png?alt=media&amp;token=0622a329-2a4d-4f45-ab09-1553c4179bd4" alt=""><figcaption></figcaption></figure>

To view the results on IPFS, navigate to the IPFS CID result output.

{% hint style="info" %}
Please be patient! IPFS can take some time to propagate and doesn't always work immediately.
{% endhint %}

```bash
https://ipfs.io/ipfs/QmZ9jvBxfwdeHTXpvDWnTqkLLWvdANzkDstph2EnXMfU4w
```

{% embed url="<https://drive.google.com/file/d/19_oEBcpKSe1n91AyGdIePJdQMzz1KjoD/view?usp=drive_link>" %}
**Results of SDV job on Output Directory**
{% endembed %}

As Lilypad modules are currently deterministic, running this command with the same text prompt will produce the same image, since the same seed is also used (the default seed is 0).

## Tuning an output

If you wish to specify more than one tunable, such as the number of steps, simply add more `-i` flags. For example, to change or improve the quality of the image generated add "Steps=x" with x = (min: 5. Max: 200):

```bash
lilypad run sdv-pipeline -i Prompt="an astronaut floating against a white background" -i Steps=69
```

{% hint style="info" %}
See the [options and tunables](#options-and-tunables) section for more information on what tunables are available.
{% endhint %}

{% hint style="info" %}
See this [beginner-friendly article](https://aituts.com/stable-diffusion-seed/) on how seeds work for more info on this
{% endhint %}

## Options and tunables

The following tunables are available. All of them are optional, and have default settings that will be used if you do not provide them.

| Name             | Description                                    | Default                           | Available options                                                                                                                                                                                                                                                                                         |
| ---------------- | ---------------------------------------------- | --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `Prompt`         | A text prompt for the model                    | "question mark floating in space" | Any string                                                                                                                                                                                                                                                                                                |
| `Framerate`      | The framerate of the video                     | 8                                 | Any valid positive integer between 1 and 20                                                                                                                                                                                                                                                               |
| `Seed`           | A seed for the image model                     | 42                                | Any valid non-negative integer                                                                                                                                                                                                                                                                            |
| `Steps`          | The number of steps to run the model for       | 50                                | Any valid non-negative integer from 5 to 200 inclusive                                                                                                                                                                                                                                                    |
| `Scheduler`      | The scheduler to use for the model             | `normal`                          | `normal`, `karras`, `exponential`, `sgm_uniform`, `simple`, `ddim_uniform`                                                                                                                                                                                                                                |
| `Sampler`        | The sampler to use for the model               | `euler_ancestral`                 | `"euler"`, `"euler_ancestral"`, `"heun"`, `"heunpp2"`, `"dpm_2"`, `"dpm_2_ancestral"`, `"lms"`, `"dpm_fast"`, `"dpm_adaptive"`, `"dpmpp_2s_ancestral"`, `"dpmpp_sde"`, `"dpmpp_sde_gpu"`, `"dpmpp_2m"`, `"dpmpp_2m_sde"`, `"dpmpp_2m_sde_gpu"`, `"dpmpp_3m_sde"`, `"dpmpp_3m_sde_gpu"`, `"ddpm"`, `"lcm"` |
| `Batching`       | How many images to produce                     | `1`                               | `1`, `2`, `4`, `8`                                                                                                                                                                                                                                                                                        |
| `VideoSeed`      | A seed for the video model                     | 42                                | Any valid non-negative integer                                                                                                                                                                                                                                                                            |
| `VideoSteps`     | The number of steps to run the video model for | 50                                | Any valid non-negative integer from 5 to 70 inclusive                                                                                                                                                                                                                                                     |
| `VideoScheduler` | The scheduler to use for the video model       | `normal`                          | `normal`, `karras`, `exponential`, `sgm_uniform`, `simple`, `ddim_uniform`                                                                                                                                                                                                                                |
| `VideoSampler`   | The sampler to use for the video model         | `euler_ancestral`                 | `"euler"`, `"euler_ancestral"`, `"heun"`, `"heunpp2"`, `"dpm_2"`, `"dpm_2_ancestral"`, `"lms"`, `"dpm_fast"`, `"dpm_adaptive"`, `"dpmpp_2s_ancestral"`, `"dpmpp_sde"`, `"dpmpp_sde_gpu"`, `"dpmpp_2m"`, `"dpmpp_2m_sde"`, `"dpmpp_2m_sde_gpu"`, `"dpmpp_3m_sde"`, `"dpmpp_3m_sde_gpu"`, `"ddpm"`, `"lcm"` |

Learn more about this Lilypad module on [Github](https://github.com/Lilypad-Tech/lilypad-module-sdv-pipeline).


# Llama LLM

Ollama Pipeline modules for Lilypad

## Overview

Based on Ollama, the Ollama Pipeline modules for Lilypad allow you generate text on Lilypad using various models.

### Available models

* #### Llama3

## Llama3

Llama3 is a machine learning model used for natural language processing. It is based on a transformer architecture, which enables it to handle tasks like text generation, summarization, translation, and more. When integrated with Lilypad, it leverages the platform's capabilities to provide efficient text processing for various applications.

### Prerequisites

Before running the Ollama Pipeline, make sure you have the [Lilypad CLI installed](https://docs.lilypad.tech/lilypad/lilypad-milky-way-testnet/install-run-requirements) on your machine and your private key environment variable is set. This is necessary for operations within the Lilypad network.

```bash
export WEB3_PRIVATE_KEY=<YOUR_PRIVATE_KEY>
```

### Usage

The Ollama Pipeline in Lilypad can be run using the Lilypad CLI or Docker. Below are the instructions for both of those options.

#### Lilypad

To run Ollama Pipeline using the Lilypad CLI, you can use the following command:

```bash
lilypad run ollama-pipeline:llama3-8b-lilypad2 -i Prompt='Something awesome this way comes'
```

#### Docker

To run this module in Docker, you can use the following commands:

```bash
docker run -ti --gpus all \
    -v $PWD/outputs:/outputs \
    -e PROMPT="Something awesome this way comes" \
    zorlin/ollama:llama3-8b-lilypad2
```

### Links

* [Lilypad Ollama Pipeline modules README](https://github.com/Lilypad-Tech/lilypad-module-ollama-pipeline/blob/main/README.md)
* [Source code](https://github.com/Lilypad-Tech/lilypad-module-ollama-pipeline)


# AlphaFold2

Coming soon...


# JS CLI Wrapper

Run the Lilypad CLI wrapper locally

The Lilypad CLI wrapper can run locally to create an API endpoint for running jobs on the Lilypad network. This gives developers full control to build a decentralized system running jobs on Lilypad. Github repo can be found [here](https://github.com/Lilypad-Tech/js-cli-wrapper).

Build a [front end ](https://blog.lilypadnetwork.org/setting-up-your-lilypad-front-end)or AI agent workflow that uses this API endpoint for running jobs! For inspiration, check out [this](https://github.com/Lilypad-Tech/js-cli-wrapper/tree/main/examples) JS CLI wrapper + Gradio example. Spin up a Webui with Gradio and use the api with a frontend!

**Note:** This is a beta tool and would mostly be expected to run locally. When implementing this tool, note that the POST request includes the user's Web3 private key. Stay tuned for a hosted API from Lilypad that will supplement this local CLI Wrapper.

## Getting Started

### Prerequisites

#### Installing and Setting up Lilypad Binary

1. Build the Lilypad binary:

```
git clone https://github.com/Lilypad-Tech/lilypad
cd lilypad
go build -v -o lilypad

# For Linux: Move to /usr/bin
sudo mv lilypad /usr/bin/

# For Mac: Move to /usr/local/bin
sudo mv lilypad /usr/local/bin
```

### Usage

Run `node src/index.js` to create a local endpoint using the js wrapper with either `src/run.js` or `src/stream.js`, then send a post request containing json with your funded `WEB3_PRIVATE_KEY` key set, see the quick start for more on [Quickstart](/lilypad/quickstart).

In `inputs`, each input must be preceded by the `-i` flag, including tunables. For example: `"inputs": "-i Prompt='an astronaut floating against a white background' -i Steps=50"`

**Note:** This tool is for demonstration purposes and can be used to run jobs on Lilypad for free. The tooling will be improved upon in the coming weeks for greater scalability. Use the following post request with the WEB3\_PRIVATE\_KEY below to run jobs on Lilypad. The wallet/private key below is funded with testnet tokens only and has been setup to simplify the use of this developer tool.

The endpoint can then be tested using curl

```
curl -X POST http://localhost:3000 \
-H "Content-Type: application/json" \
-d '{"pk": "'"your-private-key"'", "module": "github.com/lilypad-tech/lilypad-module-lilysay:0.1.0", "inputs": "-i Message=test"}'
```

```bash
// run.js
const { run } = require("./")

run(
  "private-key",
  "module name"
  '-i payload (key=value)'
).then((res) => {
  console.log(res)
})
```

```bash
// stream.js
const { stream } = require("./")

stream(
  "private-key",
  "module name"
  '-i payload (key=value)'
  { stream: true },
).then(() => {
  console.log("Result in ./output/result")
})
```

```bash
// index.js
const fetch = require("node-fetch")
const fs = require("fs")

const URL = "http://js-cli-wrapper.lilypad.tech"
const METHOD = "POST"
const HEADERS = {
  Accept: "application/json",
  "Content-Type": "application/json",
}
const OUTPUT = "./output"

function stream(pk, module, inputs, opts) {
  const body = JSON.stringify({ pk, module, inputs, opts })

  return fetch(URL, {
    headers: HEADERS,
    method: METHOD,
    body,
  }).then(function (res) {
    const fileStream = fs.createWriteStream(`./${OUTPUT}/result`)
    res.body.pipe(fileStream)
    res.body.on("error", (error) => {
      return { error }
    })
    fileStream.on("finish", () => {
      return { status: "done" }
    })
  })
}

function run(pk, module, inputs) {
  const body = JSON.stringify({ pk, module, inputs })

  return fetch(URL, {
    headers: HEADERS,
    method: METHOD,
    body,
  }).then((raw) => raw.json())
}

module.exports = { run, stream }
```


# Developer Tutorials

A collection of Blogs and Video's for developers :)

## Guides

<table data-view="cards"><thead><tr><th></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td>Building RAG Agents on Lilypad</td><td><a href="https://www.youtube.com/watch?v=I-_9yO4wdig">https://www.youtube.com/watch?v=I-_9yO4wdig</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FwkhCkTokdmRfsElSbUSB%2FLilypad_%20Pitch%20Deck%20(Strategic%20Round).png?alt=media&amp;token=c02e0b3f-512d-434c-83c1-c5887bfce863">Lilypad_ Pitch Deck (Strategic Round).png</a></td></tr><tr><td>Can an AI Agent Solve the Crypto UI Problem?</td><td><a href="https://narb.hashnode.dev/creating-an-ai-agent-to-help-solve-the-crypto-ux-problem">https://narb.hashnode.dev/creating-an-ai-agent-to-help-solve-the-crypto-ux-problem</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FhmJ3QnQV4oy8g7ymRJbW%2Fimage.png?alt=media&amp;token=b84d6a73-b672-4d60-b5fe-9d29900cae3b">image.png</a></td></tr><tr><td>Build a chatbot on Lilypad</td><td><a href="https://blog.lilypadnetwork.org/build-a-chatbot-with-lilypad">https://blog.lilypadnetwork.org/build-a-chatbot-with-lilypad</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FJsh2T68Sf7QfZKlbfkUu%2Fa1e7a017-32b2-4e37-bdff-e8e6e7b9fffa.webp?alt=media&amp;token=78b11490-4e70-4082-bb09-fba387e65d05">a1e7a017-32b2-4e37-bdff-e8e6e7b9fffa.webp</a></td></tr><tr><td>Websearch API powered by the Lilypad Inference API</td><td><a href="https://github.com/narbs91/lilypad-websearch-agent-demo">https://github.com/narbs91/lilypad-websearch-agent-demo</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F60dXy9xCu6wNtjl4Dngw%2FScreenshot%202025-04-14%20at%203.10.41%E2%80%AFPM.png?alt=media&amp;token=fea50fa2-6f5e-4570-aa08-b8008020083e">Screenshot 2025-04-14 at 3.10.41 PM.png</a></td></tr><tr><td>Module Builder guide</td><td><a href="https://blog.lilypadnetwork.org/lilypad-module-builder-guide">https://blog.lilypadnetwork.org/lilypad-module-builder-guide</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fuf7iDGJI4FmdEtRqv26A%2F3666cafc-9eae-47bf-b8e3-7fcc8257a527.webp?alt=media&amp;token=bc605a6c-9410-4a68-89b4-86672933b136">3666cafc-9eae-47bf-b8e3-7fcc8257a527.webp</a></td></tr><tr><td>Build a Open-source Perplexity Clone</td><td><a href="https://narb.hashnode.dev/create-a-perplexity-clone-powered-by-open-source-technology">https://narb.hashnode.dev/create-a-perplexity-clone-powered-by-open-source-technology</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FHWH8imlEPh2nyQwgJMnu%2F0f06d72e-b16e-4d9f-bc9a-2284f618f078.webp?alt=media&amp;token=61434302-446e-4df0-ba7d-e736fdd4709d">0f06d72e-b16e-4d9f-bc9a-2284f618f078.webp</a></td></tr><tr><td>Lilypad blog</td><td><a href="https://blog.lilypadnetwork.org/">https://blog.lilypadnetwork.org/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FHJ9kkA42uWrYkQZjrhIp%2FGitbook%20Banner%20(970%20x%20365%20px).svg?alt=media&amp;token=2921af1b-9b60-4929-8c82-a75f9920e859">Gitbook Banner (970 x 365 px).svg</a></td></tr></tbody></table>

## Video Tutorials

<table data-view="cards"><thead><tr><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>Build a Lilypad Module</td><td><a href="https://www.youtube.com/watch?v=JXERw3Idwks">https://www.youtube.com/watch?v=JXERw3Idwks</a></td></tr><tr><td>Run Lilypad Modules</td><td><a href="https://www.youtube.com/watch?v=vzHIXjFNeXg">https://www.youtube.com/watch?v=vzHIXjFNeXg</a></td></tr><tr><td>CLI Wrapper Frontend</td><td><a href="https://www.youtube.com/watch?v=vzHIXjFNeXg">https://www.youtube.com/watch?v=vzHIXjFNeXg</a></td></tr><tr><td>Encode x Filecoin Open Data Hack Video 2023</td><td><a href="https://www.youtube.com/watch?v=fJGTUg39n-0">https://www.youtube.com/watch?v=fJGTUg39n-0</a></td></tr><tr><td>Lilypad Youtube</td><td><a href="https://www.youtube.com/@lilypadnetwork">https://www.youtube.com/@lilypadnetwork</a></td></tr></tbody></table>

## Helpful Resources

### AI Resources

* [RAG vs Finetuning](https://www.rungalileo.io/blog/optimizing-llm-performance-rag-vs-finetune-vs-both)
* [The promise and challenges of crypto + AI ](https://vitalik.eth.limo/general/2024/01/30/cryptoai.html)- Vitalik Buterin&#x20;
* [Data on Notable AI Models](https://epoch.ai/data/notable-ai-models)
* [Towards decentralized AI](https://www.firstbatch.xyz/blog/towards-decentralized-ai-part-1-data-collection)
* [Understanding the Intersection of Crypto and AI ](https://www.galaxy.com/insights/research/understanding-intersection-crypto-ai/)- Galaxy Digital Research

### GPU Info

* [How do graphics cards work?](https://www.youtube.com/watch?v=h9Z4oGN89MU)
* Nvidia CUDA [Toolkit](https://developer.nvidia.com/cuda-toolkit)
* [Nvidia CUDA in 100 Seconds](https://www.youtube.com/watch?v=pPStdjuYzSI)

### Bacalhau Resources

* [Bacalhau.org](https://www.bacalhau.org/)
* [Impact of Compute over Data](https://www.youtube.com/watch?v=jSzb7q00_0c) - Juan Benet
* [CoD vision & goals](https://www.youtube.com/watch?v=-d4iJm-RbyA) - Juan Benet
* [Lessons for building Distributed Compute](https://www.youtube.com/watch?v=GHGYHuIboy4) - Juan Benet & Molly Mackinlay
* [State of CoD 2023](https://www.youtube.com/watch?v=w-cRNst68L0) - David Aronchick
* [Containers at the Edge](https://softwareengineeringdaily.com/2024/08/22/containers-at-the-edge-with-david-aronchick/) - David Aronchick&#x20;


# Helpful Resources

Useful links on topics that relate to the Lilypad Network

## AI Resources

* [RAG vs Finetuning](https://www.rungalileo.io/blog/optimizing-llm-performance-rag-vs-finetune-vs-both)
* [The promise and challenges of crypto + AI ](https://vitalik.eth.limo/general/2024/01/30/cryptoai.html)- Vitalik Buterin&#x20;
* [Data on Notable AI Models](https://epoch.ai/data/notable-ai-models)
* [Towards decentralized AI](https://www.firstbatch.xyz/blog/towards-decentralized-ai-part-1-data-collection)
* [Understanding the Intersection of Crypto and AI ](https://www.galaxy.com/insights/research/understanding-intersection-crypto-ai/)- Galaxy Digital Research

## GPU Info

* [How do graphics cards work?](https://www.youtube.com/watch?v=h9Z4oGN89MU)
* Nvidia CUDA [Toolkit](https://developer.nvidia.com/cuda-toolkit)
* [Nvidia CUDA in 100 Seconds](https://www.youtube.com/watch?v=pPStdjuYzSI)

## Bacalhau Resources

* [Bacalhau.org](https://www.bacalhau.org/)
* [Impact of Compute over Data](https://www.youtube.com/watch?v=jSzb7q00_0c) - Juan Benet
* [CoD vision & goals](https://www.youtube.com/watch?v=-d4iJm-RbyA) - Juan Benet
* [Lessons for building Distributed Compute](https://www.youtube.com/watch?v=GHGYHuIboy4) - Juan Benet & Molly Mackinlay
* [State of CoD 2023](https://www.youtube.com/watch?v=w-cRNst68L0) - David Aronchick
* [Containers at the Edge](https://softwareengineeringdaily.com/2024/08/22/containers-at-the-edge-with-david-aronchick/) - David Aronchick&#x20;

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-cover data-type="files"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>Lilypad Blog</td><td><a href="https://blog.lilypadnetwork.org/">View blog</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fv6LeXxAuc7k0dk5CqqOk%2FScreenshot%202024-05-17%20at%2011.14.42%E2%80%AFAM.png?alt=media&amp;token=50aca1ae-ed1e-49c1-a52c-9f7caa9f0444">Screenshot 2024-05-17 at 11.14.42 AM.png</a></td><td><a href="https://blog.lilypadnetwork.org/">https://blog.lilypadnetwork.org/</a></td></tr><tr><td>Lilypad YouTube</td><td><a href="https://www.youtube.com/@lilypadnetwork">Watch videos</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-5533ed78e232fb1c33af71ee6f9bff1b1913ad67%2F8gzcr6RpGStvZFA2qRt4v6.jpg?alt=media">8gzcr6RpGStvZFA2qRt4v6.jpg</a></td><td></td></tr><tr><td>Encode x Filecoin Open Data Hack Video [Sept 23]</td><td><a href="https://www.youtube.com/watch?v=fJGTUg39n-0">Watch video</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-da015ee0ca74cff830b26d014c1be22c189f9409%2FAlly_Haire_Mentor.png?alt=media">Ally_Haire_Mentor.png</a></td><td><a href="https://www.youtube.com/watch?v=fJGTUg39n-0">https://www.youtube.com/watch?v=fJGTUg39n-0</a></td></tr><tr><td>Encode x Filecoin Open Data Hack Slides [Sept 23]</td><td><a href="https://drive.google.com/file/d/15Mk06cL-G1rpkm1iIDi-zJe0DgqBUzeQ/view?usp=sharing">View slide</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-78eef13ce8f2a3bdc93561b4b29ae22817dd082f%2FOpen%20Data%20Hack.jpg?alt=media">Open Data Hack.jpg</a></td><td><a href="https://drive.google.com/file/d/15Mk06cL-G1rpkm1iIDi-zJe0DgqBUzeQ/view?usp=sharing">https://drive.google.com/file/d/15Mk06cL-G1rpkm1iIDi-zJe0DgqBUzeQ/view?usp=sharing</a></td></tr></tbody></table>


# Ecosystem Projects

Projects built with Lilypad

Explore real-world applications leveraging the Lilypad Network!

## End User Apps

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-cover data-type="files"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>ai learning buddy</td><td>A learning assistant with chat capabilities, note management, and retrieval-augmented generation (RAG) technology powered by Lilypad Anura API</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FCz8nkEPsZd5gSERgLVKb%2Foutput.png?alt=media&amp;token=65496cc0-9b45-4759-8843-48de2fbd0528">output.png</a></td><td><a href="https://github.com/jamiebones/ai-learning-buddy">https://github.com/jamiebones/ai-learning-buddy</a></td></tr><tr><td>Lilypad Anura Playground</td><td>Run any model hosted on the Lilypad network in a simple chat interface.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FE8LAfkfrs7XO1RrP8nMJ%2FScreenshot%202025-04-08%20at%204.06.35%E2%80%AFPM.png?alt=media&amp;token=6e76a6af-a503-4c68-af1a-c87923a50f7e">Screenshot 2025-04-08 at 4.06.35 PM.png</a></td><td><a href="https://anura-playground.vercel.app/">https://anura-playground.vercel.app/</a></td></tr><tr><td><a href="https://www.youtube.com/watch?v=5Hq3lUobrN4&#x26;t=5s">AI Oncologist Agent</a></td><td>A research agent template that reads papers, aggregates relevant info into reports, and answers questions about research.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F4q45xLQnH0u2eUxGr7Sc%2Foutput.png?alt=media&amp;token=256677b0-d1dc-486c-9ad1-db4cb9bc60a7">output.png</a></td><td><a href="https://github.com/mavericb/ai-oncologist">https://github.com/mavericb/ai-oncologist</a></td></tr><tr><td>Customer Support agent</td><td>A customer support agent with chat capabilities that uises text input from docs, blogs, etc. to provide answers on specific questions.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FP27MssZhHoNvqbBmbueq%2Foutput.png?alt=media&amp;token=7ded0e98-e0fb-4b07-a2d7-b73a5f622455">output.png</a></td><td><a href="https://docs.lilypad.tech/lilypad/use-cases-agents-and-projects/agents/rag-support-agent">https://docs.lilypad.tech/lilypad/use-cases-agents-and-projects/agents/rag-support-agent</a></td></tr><tr><td>WebSearch UI</td><td>Combines websearch and AI powered answers using the Lilypad Inference API to help answer questions based on the latest happenings provided as context.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F60dXy9xCu6wNtjl4Dngw%2FScreenshot%202025-04-14%20at%203.10.41%E2%80%AFPM.png?alt=media&amp;token=fea50fa2-6f5e-4570-aa08-b8008020083e">Screenshot 2025-04-14 at 3.10.41 PM.png</a></td><td></td></tr><tr><td>Gradio x Lilypad Interface</td><td>Run AI jobs from Gradio interface.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FLWX7ueNSqdKxYaRakKr7%2Fgradio.png?alt=media&amp;token=d32f5359-22af-47cf-aff3-9a522d22f414">gradio.png</a></td><td><a href="https://github.com/Lilypad-Tech/js-cli-wrapper/tree/main/examples/gradio-react-demo">https://github.com/Lilypad-Tech/js-cli-wrapper/tree/main/examples/gradio-react-demo</a></td></tr><tr><td>Lilytree</td><td>Stylized tree art from ETH addresses.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FH2Fi7GhZk3EyRKnmY0Ub%2Flilytree.jpeg?alt=media&amp;token=ce59f31a-b763-4d58-86c1-68dc17212f00">lilytree.jpeg</a></td><td><a href="https://github.com/rhochmayr/lilypad-module-lilytree">https://github.com/rhochmayr/lilypad-module-lilytree</a></td></tr><tr><td>Sir Croaksworth's Roast dApp</td><td>Get your portfolio roasted.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F2eFzFqLIJkcGYaidKBii%2FScreenshot%202025-04-01%20at%205.56.39%E2%80%AFPM.png?alt=media&amp;token=99c60bcd-93f4-49c9-b918-02545f460482">Screenshot 2025-04-01 at 5.56.39 PM.png</a></td><td><a href="https://roast-app-delta.vercel.app/">https://roast-app-delta.vercel.app/</a></td></tr><tr><td>SDXL Turbo UI</td><td>Generate vibrant AI art from a text prompt.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FkYzqtSP4xmgimu5aP6Q8%2Fgenerated-image.png?alt=media&amp;token=77c383be-8a53-4f98-b9b2-fada4842fb48">generated-image.png</a></td><td><a href="https://lp-sdxl-ui.vercel.app/">https://lp-sdxl-ui.vercel.app/</a></td></tr><tr><td>the great debAIte</td><td>AI personas debate your topic.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FBI0wVEMgqQMZCaURTsON%2FScreenshot%202025-04-02%20at%209.38.49%E2%80%AFPM.png?alt=media&amp;token=7b87eb7b-a8b0-4abe-800e-3d312c6bbdd8">Screenshot 2025-04-02 at 9.38.49 PM.png</a></td><td><a href="https://debaite-sigma.vercel.app/">https://debaite-sigma.vercel.app/</a></td></tr><tr><td>Waterlily.ai</td><td>Ethical generative AI-Art DApp.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F2O3OAjrRUjngCytbR9i6%2Fwaterlily.png?alt=media&amp;token=583b27e9-d687-482b-8092-95a988cf988d">waterlily.png</a></td><td><a href="/lilypad/use-cases-agents-and-projects/project-showcase/waterlily.ai">Waterlily.ai</a></td></tr><tr><td>ExtraLabs x Lilypad</td><td>A work-in-progress proof of concept utilizing Extra Labs and Lilypad to create a Lilypad module for running multiple geospatial jobs such as 3d modeling.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fw3MQeUaNR4G7x99l3jfA%2Foutput.png?alt=media&amp;token=ef550189-c5c1-47d3-aaa4-133e14d0f32e">output.png</a></td><td><a href="https://docs.lilypad.tech/lilypad/use-cases-agents-and-projects/project-showcase/extra-labs-x-lilypad">https://docs.lilypad.tech/lilypad/use-cases-agents-and-projects/project-showcase/extra-labs-x-lilypad</a></td></tr></tbody></table>

**EthGlobal HackFS - May-June '24**

<table data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-cover data-type="files"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><strong>CipherCraft</strong></td><td>A decentralized hub for federated model training on access controlled private datasets.</td><td><a href="https://github.com/Shubham-Rasal/CipherCraft">Source code</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fl9E91zZpiexBDftdzTvK%2Fciphercraft.webp?alt=media&amp;token=e718a60c-7ff4-496c-ac26-c9319bb93f96">ciphercraft.webp</a></td><td><a href="https://ethglobal.com/showcase/ciphercraft-3g8kp">https://ethglobal.com/showcase/ciphercraft-3g8kp</a></td></tr><tr><td><strong>NexTown</strong></td><td>A decentralized AI computation and storage platform allowing users to run AI models via LilyPad, encrypt results with their public key, and securely store outputs on Lighthouse, accessible only by the user’s private key.</td><td><a href="https://github.com/DogukanGun/hackfs24-ai-marketplace">Source code</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-5966bea37656e7f38e58df1fdc3cad97b5ac1f07%2Fnextown.png?alt=media">nextown.png</a></td><td><a href="https://ethglobal.com/showcase/nextown-4fpk9">https://ethglobal.com/showcase/nextown-4fpk9</a></td></tr><tr><td><strong>Filecoin ML Engine</strong></td><td>Onchain ML engine to create dApp ML agents on top of Filecoin.</td><td><a href="https://github.com/Prajjawalk/filecoin-ML-engine">Source code</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FT2JUGwmeR8HmqiIGktiC%2Ffilecoin-ml-engine.png?alt=media&amp;token=4b7860a9-d013-4b96-9592-dd3e197cd5c7">filecoin-ml-engine.png</a></td><td><a href="https://ethglobal.com/showcase/filecoin-ml-engine-x1z4q">https://ethglobal.com/showcase/filecoin-ml-engine-x1z4q</a></td></tr><tr><td><strong>Dogtoken ai erc404</strong></td><td>ERC404 token with custom Lilypad stable diffusion to generate unique images for holders.</td><td><a href="https://github.com/lucasespinosa28/dogtoken">Source code</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FcRJSQZylfuctcPMJ8Wi1%2Fdogtoken.png?alt=media&amp;token=cf1405b2-cc30-4f30-af7c-96c4e11ccccf">dogtoken.png</a></td><td><a href="https://ethglobal.com/showcase/dogtoken-ai-erc404-xtjh5">https://ethglobal.com/showcase/dogtoken-ai-erc404-xtjh5</a></td></tr><tr><td><strong>Rock Paper Ninja</strong></td><td>A web3 take on Rock Paper Scissors, utilizing Lilypad to create generative avatars.</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FQ2OPb7DgIuaN7sl1of3h%2Frockpaperninja.png?alt=media&amp;token=34b35787-8f69-4d8e-a529-6b150aa069f9">rockpaperninja.png</a></td><td><a href="https://ethglobal.com/showcase/rock-paper-ninja-w40wj">https://ethglobal.com/showcase/rock-paper-ninja-w40wj</a></td></tr></tbody></table>

**EthGlobal Sydney - May '24**

<table data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td></td><td><strong>Learnr Idle Mining</strong></td><td>An innovative NFT game where characters evolve and improve their tasks over time.<br><br><a href="https://github.com/Taresin/Learnr">Source code</a></td><td><a href="https://ethglobal.com/showcase/learnr-idle-mining-7pscw">https://ethglobal.com/showcase/learnr-idle-mining-7pscw</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FvxaSB7EWmaqkx8CQ08fE%2Flearnr.webp?alt=media&amp;token=9418fc0b-52e1-4eb1-a613-6bc26b47dd30">learnr.webp</a></td></tr><tr><td></td><td><strong>SlickCI</strong><br><br>One click deployment of CI/CD workflows on to decentralised cloud infrastructure.<br><br><a href="https://github.com/shadowfaz16/slick-ci">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/slickci-h5a0y">https://ethglobal.com/showcase/slickci-h5a0y</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fxg4Jt0RVG60u67aEVW0y%2Fci.webp?alt=media&amp;token=82bda10b-d249-475b-b6e9-972bc27101d8">ci.webp</a></td></tr><tr><td><br><strong>On-Chain Personality  - Honorable mention</strong></td><td></td><td>Leverages on-chain Ethereum mainnet transactions to classify wallets into six predefined buckets/sub-ecosystems: Airdrop Farmer, NFT Collectooor, Snipoor, Shitcoinooor, EVM Enjoyoor, and NPCs.</td><td><a href="https://ethglobal.com/showcase/on-chain-personality-5nir5">https://ethglobal.com/showcase/on-chain-personality-5nir5</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F7hocImqhtMhGaz2C5HvY%2Focp.webp?alt=media&amp;token=1f0b02a9-9faa-4b47-9972-a6a079e66591">ocp.webp</a></td></tr><tr><td><strong>Privacy Avengers - Could use Lilypad tech</strong><br></td><td>A protocol for verifiable machine unlearning, solving AI privacy issues by securely and efficiently detecting and removing user data.<br></td><td><a href="https://github.com/Mingzhe-W/privacy_avengers">Source code</a></td><td><a href="https://ethglobal.com/showcase/privacy-avengers-9yc3i">https://ethglobal.com/showcase/privacy-avengers-9yc3i</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FrHv6ogk8QfKVpFXcXtYT%2Fdav.webp?alt=media&amp;token=dd94970c-c45f-481f-9b8d-cc2303ac8236">dav.webp</a></td></tr></tbody></table>

**EthGlobal Istanbul - November '23**

<table data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-cover data-type="files"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><strong>dub3.ai</strong><br><br>dub3.ai is an innovative marketplace harnessing the power of Cartesi, enabling users to transform their voice into unique NFTs and generate text-to-speech content using cutting-edge AI technology.<br><br><a href="https://github.com/dub3-space/dub3-hackthon-project">Source code</a></td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FZK0D3WFJvoWfrXIVkm8b%2Fdub3.webp?alt=media&amp;token=d5d54fdb-7dc9-43b2-844d-2b29eadce449">dub3.webp</a></td><td><a href="https://ethglobal.com/showcase/dub3-ai-h1riu">https://ethglobal.com/showcase/dub3-ai-h1riu</a></td></tr><tr><td><strong>HuggingLily</strong></td><td>Run HuggingFace pipelines via dashboard with just one click.</td><td><br><a href="https://github.com/tolgayayci/ethglobal-istanbul">Source code</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F0kxXNpfTrizzBOW0hzpJ%2Fhugginglily.webp?alt=media&amp;token=4a900dc0-b9bf-4a7d-880f-2847296a8833">hugginglily.webp</a></td><td><a href="https://ethglobal.com/showcase/hugginglily-wvusa">https://ethglobal.com/showcase/hugginglily-wvusa</a></td></tr></tbody></table>

**ETHOnline - October '23**

<table data-view="cards"><thead><tr><th></th><th></th><th><select></select></th><th data-hidden data-type="content-ref"></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td><strong>Lilypad-Kamu</strong></td><td>A custom lilypad module for SQL streaming through kamu.dev, this provides provenance, version control, and provenance of derivative datasets.<br>Integrates the desci project kamu.dev into a bacalhau job to be consumed by lilypad v2.<br><br><a href="https://github.com/polus-arcticus/lilypad-module-kamu/blob/main/lilypad_module.json.tmpl">Source code</a></td><td></td><td><a href="https://github.com/polus-arcticus/lilypad-module-kamu/blob/main/lilypad_module.json.tmpl">https://github.com/polus-arcticus/lilypad-module-kamu/blob/main/lilypad_module.json.tmpl</a></td><td><a href="https://github.com/polus-arcticus/lilypad-module-kamu/blob/main/lilypad_module.json.tmpl">https://github.com/polus-arcticus/lilypad-module-kamu/blob/main/lilypad_module.json.tmpl</a></td><td>This project puts kamu.dev into a bacalhau container and runs it by calling a lilypad v2 from the cli. It demonstrates that sql streaming can be performed through a lilypad module to help mange the provenance and veracity of a living dataset pulling from theoretically multiple sources. It takes a manifest from an external api, pulls it into kamu and than pushes it as a dag-pb to ipfs.</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F7cfZlUMqFntSDXREJ3Zm%2Flilypad-kamu.png?alt=media&amp;token=29f46a0b-80bd-4b49-9cb0-6fd870a46311">lilypad-kamu.png</a></td></tr><tr><td><strong>Project C</strong></td><td>Dive into the cosmic engineering that propels your personalized galaxy of knowledge.<br><br><br><a href="https://github.com/0xgoldenlion/project-C">Source code</a></td><td></td><td><a href="https://github.com/0xgoldenlion/project-C">https://github.com/0xgoldenlion/project-C</a></td><td><a href="https://ethglobal.com/showcase/project-c-y6m3x">https://ethglobal.com/showcase/project-c-y6m3x</a></td><td><p>"In the infinite expanse of space, stars, and galaxies, Project C is your personal constellation, lighting the path to knowledge that speaks directly to your curiosity and growth.</p><h4>""Economic Supernovas: The Integration of Web3 for Payments""</h4><p>In the financial cosmos, web3 technology fuels the economic engines of Project C. It’s the supernova that disrupts the traditional transaction space-time continuum, allowing a seamless flow of payments across the galaxy. Your journey is no longer tethered to the old ways of monetary exchange; instead, you soar through premium learning experiences with the ease of a shooting star.</p><h4>""Artificial Intelligence, Natural Learning: The OpenAI API and lilypad Cosmos""</h4><p>In the heart of Project C lies a pulsating core of OpenAI's sophisticated language model and and lilypad decentralized infrastructure . This isn’t just a tool; it’s your AI co-pilot, navigating the learning material and adapting it in real-time to suit your educational needs. It’s an odyssey that personalizes, empathizes, and grows with you, a sentient guide in your exploration of the unknown realms of knowledge.</p><h4>""The ORM Galaxy: Streamlining the Stellar Database Interactions""</h4><p>What makes Project C a universe rather than a solitary planet is its intricate network of data interactions, all streamlined with the efficiency of ORMs. This system acts as the gravitational force that holds the learning modules together, allowing for an organized, accessible flow of information. It’s the galaxy where raw data are forged into stars of insight, each one ready to explode into a supernova of understanding."</p></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FAhhyG0hX6AePzfIMDIhp%2Fproject-c.png?alt=media&amp;token=d411944e-2e67-449d-a2e7-a418a3834f41">project-c.png</a></td></tr></tbody></table>

**Open Data Hack - September '23**

<table data-view="cards"><thead><tr><th></th><th></th><th><select multiple></select></th><th data-hidden data-type="content-ref"></th><th data-hidden></th><th data-hidden data-type="files"></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td><strong>AI Capwyn</strong></td><td>ML Tooling for Security Audits. Validate and test the status of deployed Lilypad Compute Networks. The contracts ensure seamless interaction with the network, while the test script checks for network availability and contract functionality.<br><br><a href="https://github.com/jeytuan/OpenDataHackathon_Lilypad">Source code</a></td><td></td><td><a href="https://github.com/jeytuan/OpenDataHackathon_Lilypad">https://github.com/jeytuan/OpenDataHackathon_Lilypad</a></td><td>ML Tooling for Security Audits. Validate and test the status of deployed Lilypad Compute Networks. The contracts ensure seamless interaction with the network, while the test script checks for network availability and contract functionality. This foundational work aligns with AI Capwn's vision to integrate AI and blockchain on Azure, aiming to provide more secure and efficient decentralized solutions.</td><td></td><td><a href="https://github.com/jeytuan/OpenDataHackathon_Lilypad">https://github.com/jeytuan/OpenDataHackathon_Lilypad</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F1qOWRtDABNBt2q2E0fuY%2Fai-capwyn.png?alt=media&amp;token=38b535b5-a562-4367-9448-0681be838083">ai-capwyn.png</a></td></tr><tr><td><strong>Lilypad Hub</strong></td><td>This platform, much like Docker Hub, is a one-stop destination for all your module needs as a Lilypad user. Includes a feature-rich playground where you can seamlessly test job specifications, experiment and fine-tune your code in a safe environment.<br><br><a href="https://github.com/oBLAZERo2001/lilypad-hub">Source code</a></td><td></td><td><a href="https://github.com/oBLAZERo2001/lilypad-hub">https://github.com/oBLAZERo2001/lilypad-hub</a></td><td><p>Introducing a groundbreaking platform designed with the utmost convenience for programmers in mind. This platform, much like Docker Hub, is a one-stop destination for all your module needs as a Lilypad user.</p><p>Our comprehensive platform includes a feature-rich playground where you can seamlessly test job specifications. You'll find it easy to experiment and fine-tune your code in a safe environment.</p><p>But that's not all; we understand the importance of privacy and organization. That's why we offer a private module library for you to securely save and manage all your modules. No more clutter or confusion – just a clean, organized space for your work.</p><p>As if that wasn't enough, we also provide an Explore screen where you can discover battle-tested modules uploaded by fellow users. It's a treasure trove of tried-and-true solutions to common programming challenges.</p></td><td></td><td><a href="https://youtu.be/sG9EPa0h5x0">https://youtu.be/sG9EPa0h5x0</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F40yMrIAJVoyij1aHwCTb%2Flilypadhub.png?alt=media&amp;token=d03f707e-46e3-406f-b790-ba8ac64f191f">lilypadhub.png</a></td></tr><tr><td><strong>Obsidian-Lilypad</strong></td><td>A user can plug obsidian-lilypad into their obsidian instance to fetch dpids from desci labs, have chat gpt summarize it, and use that as a prompt in runSDXL in lilypad. Helia/ipfs and kubo delegated content routing is used to splay dags and fetch content from IPFS.<br><br><a href="https://github.com/polus-arcticus/obsidian-lilypad">Source code</a></td><td></td><td><a href="https://github.com/polus-arcticus/obsidian-lilypad">https://github.com/polus-arcticus/obsidian-lilypad</a></td><td></td><td></td><td><a href="https://www.youtube.com/watch?v=H4Ww6vjLGj8&#x26;ab_channel=TaylorHulsmans">https://www.youtube.com/watch?v=H4Ww6vjLGj8&#x26;ab_channel=TaylorHulsmans</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F3ZS4nR1PlLFbrdHeryHk%2Fobsidian-desci.png?alt=media&amp;token=91d7afab-a9b2-40f5-b952-6d037c9ec796">obsidian-desci.png</a></td></tr><tr><td><strong>Decenter AI</strong></td><td><p>DeCenter AI functions as a PaaS infrastructure, empowering machine learning engineers to expedite and make the training of AI models more cost-effective through decentralized parallel training methods.</p><p>The core objective of DeCenter AI is to democratize and decentralize AI model training.<br><br><a href="https://github.com/orgs/DeCenter-AI/repositories">Source code</a></p></td><td></td><td><a href="https://github.com/orgs/DeCenter-AI/repositories">https://github.com/orgs/DeCenter-AI/repositories</a></td><td>By offering a distributed training platform, it allows data scientists, machine learning engineers, researchers, and AI specialists to collaboratively contribute to the training of AI models. Structured around a distributed parallel training mechanism, DeCenter AI has been designed to facilitate the training of various ML and DL models in a significantly reduced time frame and cost compared to the current norms. Our platform incorporates a built-in incentive system, fueled by DCEN Tokens. This system not only rewards contributors and participants but also encourages them to undertake tasks such as reviewing, testing, and rating AI models.</td><td></td><td><a href="https://github.com/orgs/DeCenter-AI/repositories">https://github.com/orgs/DeCenter-AI/repositories</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FLwltagi5Os6HVqGJGNxd%2Fdecenterai.webp?alt=media&amp;token=a16ead07-e8e6-4ba2-a531-753e8dd47838">decenterai.webp</a></td></tr><tr><td><strong>Lilylatte</strong></td><td>Lilylatte is a DataDAO designed to bridge the gap between onchain and offchain data, focusing on web3 market research. Data owners, or web3 citizens, link their wallets and engage in dialogues generated by our LLM model, 'Lily.'<br><br><a href="https://github.com/Caruso33/LilyLatte_OpenDataHack">Source code</a></td><td></td><td><a href="https://github.com/Caruso33/LilyLatte_OpenDataHack">https://github.com/Caruso33/LilyLatte_OpenDataHack</a></td><td><p>These conversations dig deep into users' behaviors, enabling them to own their data while offering high-quality, crowdsourced opinions. DAO members can curate member's opinion by voting.</p><p>Market researchers can sift through these curated opinions to conduct targeted 'dataQuests'—interviews with specific users to gain even deeper insights. Lilylatte sets itself apart by allowing data owners complete control and monetization opportunities, and researchers access to genuine, highly relevant data. The long-term vision is to evolve Lily into a smart assistant that can autonomously design surveys and provide in-depth market analysis, all trained on the rich, ethically-sourced dataset from the DataDAO.</p></td><td></td><td><a href="https://www.loom.com/share/9174a6b162a240649ebb431ecd941388?sid=3f5674b0-1025-476d-8898-6a57229599de">https://www.loom.com/share/9174a6b162a240649ebb431ecd941388?sid=3f5674b0-1025-476d-8898-6a57229599de</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FQYiZtJ0mrbvTa3DnpcUH%2Flilylatte.png?alt=media&amp;token=b46600a0-4b11-4b10-a629-ad61983100d8">lilylatte.png</a></td></tr><tr><td><strong>Tiny Hops</strong></td><td>An intents computational framework for describing lilypad workflows with a serialization format that allows you to replace variables within solidity, onchain which allows you to configure the receiveResults to feedback and trigger future jobs.<br><br><a href="https://github.com/zcstarr/tiny-hops">Source code</a></td><td></td><td><a href="https://github.com/zcstarr/tiny-hops">https://github.com/zcstarr/tiny-hops</a></td><td><p>Users are able to give workflows a balance, which in turn enables the workflows to simply pay the bacalahau deposits for each new job that gets executed.</p><p>In the end I hit some issues, getting this to go end to end, but the prototype shows the direction, and the solidity test work. I think its' a great jumping off point for future work/exploration</p></td><td></td><td><a href="https://www.youtube.com/watch?v=k8QGwukfnpw">https://www.youtube.com/watch?v=k8QGwukfnpw</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F1pr1TJKrKYZy4zgvwgYA%2Ftinyhops.webp?alt=media&amp;token=d77b8b33-4381-407e-af4a-475117b37eb8">tinyhops.webp</a></td></tr><tr><td><strong>Saturn Observatory</strong></td><td><p>The goal of my project is to increase the transparency of Filecoin Saturn network – the fastest growing, community-run distributed content delivery network (CDN) for Web3.</p><p>I use historical data about the network gathered by Saturn Moonlet to prepare and publish analytical reports on Saturn Observatory website every month.<br><br><a href="https://github.com/cronian-tech/saturn-observatory">Source code</a></p></td><td></td><td><a href="https://github.com/31z4/saturn-observatory">https://github.com/31z4/saturn-observatory</a></td><td><p>The goal of my project is to increase the transparency of Filecoin Saturn network – the fastest growing, community-run distributed content delivery network (CDN) for Web3.</p><p>I use historical data about the network gathered by Saturn Moonlet (another project I built several months ago) to prepare and publish analytical reports on Saturn Observatory website (https://31z4.github.io/saturn-observatory) every month. Raw data that is used to generate the reports is available on IPFS and Filecoin.</p><p>Saturn Observatory compliments official tools like Saturn Node Dashboard (https://dashboard.saturn.tech) and Saturn Explorer (https://explorer.saturn.tech), aiming to provide better insights into the network state and performance.</p></td><td></td><td><a href="https://youtu.be/ErgUU7kwUro">https://youtu.be/ErgUU7kwUro</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fx3yjFTKAvZJ78YOsVkQW%2Fsaturn-observatory.png?alt=media&amp;token=b97f02b2-08cb-4392-a402-80744404b490">saturn-observatory.png</a></td></tr><tr><td><strong>Rejuvenate AI</strong></td><td>RejuvenateAI is the first community blockchain based project powered by community built for community to promote healthy living and achieve healthy locations where people live up to a 100 years all around the world.<br><br><a href="https://github.com/orgs/open-data-hack/repositories">Source code</a></td><td></td><td><a href="https://github.com/orgs/open-data-hack/repositories">https://github.com/orgs/open-data-hack/repositories</a></td><td><p>RejuvenateAI is the first community blockchain based project powered by community built for community to promote healthy living and achieve healthy locations where people live up to a 100 years all around the world.</p><p>So basically how it works is, we gather top nutritionists from all around the world after vetting them so they can research and provide valuable knowledge to users on our platform. So they have to apply first by uploading their credentials. If we see they are valid we can approve them using the smart contracts. Our business model works on a subscription model, so people that want to join our platform pay a subscription fee to access all our services. Services we offer include consultation services which are live meetings with the professional nutritionists, premium healthy meal and fitness plans and articles.</p><p>When users sign up we issue them NFTs, and we also collect their data like their age, height, weight, eating habits to calculate their rate of aging using AI.</p></td><td></td><td><a href="https://github.com/open-data-hack/rejuvenate">https://github.com/open-data-hack/rejuvenate</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FyGVfiZICe6hRXyuDgFtI%2Frejuvinate.webp?alt=media&amp;token=61480382-71db-4627-b360-34a1957ab903">rejuvinate.webp</a></td></tr><tr><td><strong>StoryChain</strong></td><td><p>StoryChain, a multi-level AI based story telling NFT dapp, is a new take on the ages old classic chain of stories, where different users collaboratively create stories.</p><p>Using this dapp, users create stories that have unique chapters and arts using web3, AI, NFTs, IPFS. Each page can belong to a different user.</p></td><td></td><td><a href="https://github.com/cemleme/fvm-storyChain">https://github.com/cemleme/fvm-storyChain</a></td><td><p>Once a user creates a story, they define the category (such as if it is a childrens' story), select the story AI and also the image generation AI. The system supports multiple LLMs and also different stable diffusion models, including SDXL using Lilypad CLI.</p><p>Once the story is created, users simply read the previous chapters and enter a prompt however they wish to continue.</p><p>So a story is created collaboratively by the users and each page belongs to one user with unique story and art. More if it, this page itself is minted as an NFT for the user; which can be visited on blockchain explorer or NFT Marketplaces.</p></td><td></td><td><a href="https://www.youtube.com/watch?v=AcRVj6bSHmo">https://www.youtube.com/watch?v=AcRVj6bSHmo</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FeI2ODGVkClu7iEod4GkU%2Fstorychain.webp?alt=media&amp;token=211ecf96-d610-4915-9e35-8e669a635ce5">storychain.webp</a></td></tr><tr><td><strong>Lilypad Javascript Wrapper</strong></td><td>This project provides a JavaScript wrapper to interact with the Lilypad CLI using Node.js. The wrapper utilizes the <code>node-pty</code> package to spawn a pseudo-terminal for executing the Lilypad CLI commands. It will enable the front-end developers to integrate Lilypad with other tools.<br><br><a href="https://github.com/only4sim/lilypad-javascript-wrapper">Source code</a></td><td></td><td><a href="https://github.com/only4sim/lilypad-javascript-wrapper">https://github.com/only4sim/lilypad-javascript-wrapper</a></td><td></td><td></td><td><a href="https://youtu.be/M5XvawL6kKI">https://youtu.be/M5XvawL6kKI</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FOZw1HIwfwXZchAnywkUd%2Flilypad-js-wrapper.webp?alt=media&amp;token=97d93ddd-246d-476b-8ae7-0fedea9f3027">lilypad-js-wrapper.webp</a></td></tr><tr><td><strong>Lilywrite</strong></td><td>LilyWrite combines blockchain technology and artistic expression, enabling users to craft and own unique poetic artworks. Users generate poems using credits, represented by LW tokens, and acquire them as distinctive non-fungible tokens (NFTs), each symbolizing a unique blend of creativity and expression.<br><br><a href="https://github.com/Khwahish29/lilywrite">Source code</a></td><td></td><td><a href="https://github.com/Khwahish29/lilywrite">https://github.com/Khwahish29/lilywrite</a></td><td></td><td></td><td><a href="https://github.com/Khwahish29/lilywrite">https://github.com/Khwahish29/lilywrite</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fo1utvupkyutL770TDPEg%2Flilywrite.webp?alt=media&amp;token=21210c00-42bf-4235-8eb8-7c60eaf9d315">lilywrite.webp</a></td></tr><tr><td><strong>GreenBadge</strong></td><td>GreenBadge is an innovative on-chain funding platform dedicated to fostering sustainability and a brighter future for our planet. Within our platform, organizations committed to these crucial goals will gain access to a vast network of donors and like-minded brands who share a passionate vision for the future of our environment.<br><br><a href="https://github.com/priyanshur66/greenbadge">Source code</a></td><td></td><td><a href="https://github.com/priyanshur66/greenbadge">https://github.com/priyanshur66/greenbadge</a></td><td><p>For brands, our platform offers a unique opportunity to collaborate with these organizations, not only to offset their carbon footprint but also to tap into their expertise in achieving sustainability in their own operations. As governments around the world increasingly push for environmentally responsible practices, this collaboration becomes even more vital.</p><p>What sets our platform apart is its unwavering commitment to transparency. Through the support of our generous donors, funds flow seamlessly to these organizations, creating a transparent and verifiable process. Donors, in turn, receive tangible proof of their support in the form of NFTs, which can be a</p></td><td></td><td><a href="https://youtu.be/DL6mMyURO-A">https://youtu.be/DL6mMyURO-A</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FMqNSVQrSXOLfBCsalBq7%2Fgreenbadge.webp?alt=media&amp;token=19ffdaaa-897b-4dd8-b30f-a06fde6ec020">greenbadge.webp</a></td></tr></tbody></table>

**EthGlobal Paris - July '23**

<table data-view="cards" data-full-width="false"><thead><tr><th></th><th></th><th><select multiple></select></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td><strong>Gadus CLI</strong></td><td>CLI Package for running 3D surface reconstruction with decentralized compute.<br><br><a href="https://github.com/The-Extra-Project/Gadius-CLI">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/gadus-cli-1c5qd">https://ethglobal.com/showcase/gadus-cli-1c5qd</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-8dc07be3353896b011fd9dba59883b434d5e4def%2FGadus.jpeg?alt=media">Gadus.jpeg</a></td></tr><tr><td><strong>Deehr Market</strong></td><td>Empowering Patients with ownership of their electronic health records.<br><br><a href="https://github.com/Cabal-Labs/deehr-market-client">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/deehr-market-mvpzu">https://ethglobal.com/showcase/deehr-market-mvpzu</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-c2f0a324ad5396e12dc307f298b1d35c8f74c97b%2FDeehr.png?alt=media">Deehr.png</a></td></tr><tr><td><strong>CréditDécentrale</strong></td><td>A decentralized credit score system using zkSNARKs, decentralized Computation &#x26; Storage.<br><br><a href="https://github.com/solity-research/ETHGlobalParis2023">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/creditdecentrale-wbzz4">https://ethglobal.com/showcase/creditdecentrale-wbzz4</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-f8fab52141ff611418b260e52d4c1317099907ed%2Fprojects_wbzz4_images_Screenshot_2023-07-23_at_03.00.04.png?alt=media">projects_wbzz4_images_Screenshot_2023-07-23_at_03.00.04.png</a></td></tr><tr><td><strong>Cypher Deposit</strong></td><td>Elevating financial privacy in crypto withdrawals. Securely withdraw funds via anonymous transfers &#x26; secure transactions, preserving anonymity.<br><br><a href="https://github.com/Alice-s-Deposit">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/cypher-deposit-cnhs5">https://ethglobal.com/showcase/cypher-deposit-cnhs5</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-1bb80f890c96686ddc75d10a2614e7daa0574bd8%2Fprojects_cnhs5_images_1.png?alt=media">projects_cnhs5_images_1.png</a></td></tr><tr><td><strong>LLM Bench</strong></td><td>Verifiable on-chain Large Language Models drift benchmarking inspired by <a href="https://arxiv.org/abs/2307.09009">https://arxiv.org/abs/2307.09009</a><br><br><a href="https://github.com/codethazine/llmbench">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/llmbench-hdeau">https://ethglobal.com/showcase/llmbench-hdeau</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-828b93850a7f72d55ca48bf024ffa033751a0b99%2Fprojects_hdeau_images_Schermata%202023-07-23%20alle%2000.41.28.png?alt=media">projects_hdeau_images_Schermata 2023-07-23 alle 00.41.28.png</a></td></tr><tr><td><strong>Uncensored GPT-5</strong></td><td>Uncensored 100% decentralized GPT AI chat running on the blockchain.<br><br><a href="https://ethglobal.com/showcase/uncensored-gpt-5-blockchain-15did">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/uncensored-gpt-5-blockchain-15did">https://ethglobal.com/showcase/uncensored-gpt-5-blockchain-15did</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-6644ac591f710d34e2f9026239ca2ab868f6e1c6%2Fprojects_15did_images_Screenshot%202023-07-23%20at%202.41.10%20AM.png?alt=media">projects_15did_images_Screenshot 2023-07-23 at 2.41.10 AM.png</a></td></tr><tr><td><strong>ZK Microphone</strong></td><td>Trusted audio in the age of deepfakes 🔒🎙 Generative AI is a threat to society. It enables disinformation, manipulation, and political subversion. We've built the world's first attested microphone and used ZK-SNARKs to protect authenticity and privacy.<br><br><a href="https://github.com/Miyamura80/ZKMicrophone">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/zk-microphone-8161v">https://ethglobal.com/showcase/zk-microphone-8161v</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-82a01fe290efd9ab3d2d59df5b388373b562c6b5%2Fprojects_8161v_images_DALL%C2%B7E%202023-07-23%2006.27.02%20-%20Futuristic%20looking%20karaoke%20advanced%20microphone%20in%20a%20cartoon%20style%20%2C%20filled%20with%20math%20equations.png?alt=media">projects_8161v_images_DALL·E 2023-07-23 06.27.02 - Futuristic looking karaoke advanced microphone in a cartoon style , filled with math equations.png</a></td></tr></tbody></table>

**AugmentHack Paris - July '23**

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th><select></select></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td><strong>CarpAI</strong></td><td>A fastchat LLM Inference Module for Lilypad v1.<br><br><a href="https://github.com/Xqua/CarpAI-LLM">Source code</a></td><td></td><td><a href="https://devpost.com/software/carpai-fmecgh">https://devpost.com/software/carpai-fmecgh</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-b8ad6ac739585ecb3ed62e94ccccede6066b332b%2Fleo.webp?alt=media">leo.webp</a></td></tr><tr><td><strong>Brian</strong></td><td>Decentralized AI Assistant for Web3 - Perform on-chain transactions via prompt.<br><br><a href="https://github.com/brian-knows/brian-fine-tuning">Source code</a></td><td></td><td><a href="https://devpost.com/software/brian">https://devpost.com/software/brian</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-345c7a61038e70a5ec127aaa3cafcce1db877b0a%2Fbrian.webp?alt=media">brian.webp</a></td></tr><tr><td><strong>AvaSoul</strong></td><td>Private-Model Generation &#x26; Service Platform -> Your very own unique personality in an AI Avatar (web3 character.ai)!<br><br><a href="https://github.com/mr-spaghetti-code/lilypad/tree/main">Source code</a></td><td></td><td><a href="https://devpost.com/software/avasoul-ai">https://devpost.com/software/avasoul-ai</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-8be326914d4a73eeb984e63ddaea177bf169f41e%2Fgallery.jpg?alt=media">gallery.jpg</a></td></tr><tr><td><strong>EasyCraftAI</strong></td><td>Based on academic research, EasyCraft built an AI and blockchain based supply chain management implementation matching factory capacity with customer orders for economic efficiency.<br><br><a href="https://github.com/BigTava/easycraft">Source code</a></td><td></td><td><a href="https://devpost.com/software/trusted-human-network">https://devpost.com/software/trusted-human-network</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-25eb97e05aa41e774e0e32f531cff28f6dc8a8e0%2Fgallery%20(1).jpg?alt=media">gallery (1).jpg</a></td></tr></tbody></table>

**HackFS Online - June '23**

<table data-view="cards"><thead><tr><th></th><th></th><th><select></select></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td><strong>DefiKicks</strong></td><td>A decentralized, community-governed Data DAO on Filecoin that democratizes DeFi data aggregation and TVL calculations with on-chain adapter proposals, off-chain voting &#x26; interoperable tokens.<br><br><a href="https://github.com/md0x/defikicks">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/defikicks-b1wo0">https://ethglobal.com/showcase/defikicks-b1wo0</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-1860952d4057574b4c1dbe3baebde72723a27d59%2Fprojects_b1wo0_images_android-chrome-512x512.png?alt=media">projects_b1wo0_images_android-chrome-512x512.png</a></td></tr><tr><td><strong>Decentralized Yield Data Collector</strong></td><td>A proof of concept of a decentralized yield data aggregator. It gets the pool data, does calculations on it and shows the best APY value of the reserve to the user with a push notification.<br><br><a href="https://github.com/aaytuncc/HackFS-2023">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/decentralized-yield-data-collector-hs96d">https://ethglobal.com/showcase/decentralized-yield-data-collector-hs96d</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-c7b0d75f12de7744ce086bfb67015e233750de8c%2Fprojects_hs96d_images_DYIELD%20APP.png?alt=media">projects_hs96d_images_DYIELD APP.png</a></td></tr><tr><td><strong>Pensieve</strong></td><td>Pensieve is a decentralized file storage to recreate your those moments in your memories and share with others. As you upload these memories, you can also choose to take part in the documentation of the HISTORY OF APES (HOMO SAPIENS).<br><br><a href="https://github.com/ahsueh1996/Pensieve-">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/pensieve-d0ixh">https://ethglobal.com/showcase/pensieve-d0ixh</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-da34d96cdd639cfcca592d6134d8beb430b49f32%2Fprojects_d0ixh_images_logo.png?alt=media">projects_d0ixh_images_logo.png</a></td></tr><tr><td><strong>Daggle</strong></td><td>A swiss army knife for Bacalhau, everything you need in an easily accessible dashboard.<br><br><a href="https://github.com/leostelon/daggle">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/daggle-6xot9">https://ethglobal.com/showcase/daggle-6xot9</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-450783c8f78b50e6326df41f40b4ea084b31ad4b%2Fprojects_6xot9_images_Screenshot_142.png?alt=media">projects_6xot9_images_Screenshot_142.png</a></td></tr><tr><td><strong>InferAI</strong></td><td>A single click deploy for your ML models on decentralized compute and storage powered by FVM, Bacalhau &#x26; Libp2p.</td><td></td><td><a href="https://ethglobal.com/showcase/inferai-21kyn">https://ethglobal.com/showcase/inferai-21kyn</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-51b2606b5f8e1eadc49cee9fa8827a4a07607264%2Fprojects_21kyn_images_image.png?alt=media">projects_21kyn_images_image.png</a></td></tr><tr><td><strong>Tentai</strong></td><td>An easy SDK for AI models running on decentralised computing.<br><br><a href="https://github.com/debuggingfuture/tentai">Source code</a></td><td></td><td><a href="https://ethglobal.com/showcase/tentai-ub5xn">https://ethglobal.com/showcase/tentai-ub5xn</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-d41a32556f4ea51608b462806a9de67bd27b130f%2Fprojects_ub5xn_images_TENTAI_LOGO.png?alt=media">projects_ub5xn_images_TENTAI_LOGO.png</a></td></tr><tr><td>DefiKicks</td><td>A decentralized, community-governed Data DAO on Filecoin that democratizes DeFi data aggregation and TVL calculations<br><br><a href="https://github.com/md0x/defikicks">Source code</a></td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FkUS65XQmdhOyuSfIBRhu%2Fdefikicks.png?alt=media&amp;token=92192b49-fc86-4b8a-b04f-60994552e605">defikicks.png</a></td></tr></tbody></table>


# Waterlily.ai

Waterlily.ai - An Ethical Generative AI-Art DApp

{% embed url="<https://youtu.be/oXNaCs0eyXI>" %}

[Waterlily.ai](http://waterlily.ai/) is a proof-of-concept application aiming to provide an alternative ethical solution to AI-Art generation by creating a new paradigm for AI image generation that pays royalties to the original creator and offers a new revenue stream for artists to explore.

To achieve this, Waterlily leverages the transparency & trustlessness of blockchain, with [FVM](https://fvm.dev/) smart contracts, combined with a decentralised edge compute platform in [Bacalhau](https://docs.bacalhau.org/examples/) to deliver an intuitive user experience for generative AI-Art while compensating creators for their original work with every image generation call.

Each stable diffusion model on Waterlily is trained & finetuned to a selection of artist work that has been uploaded by the artist themselves or is available in the public domain. When a user navigates to the [Waterlily.ai](http://waterlily.ai/) platform, they simply enter a text prompt for the images they want to create and choose an artist style that appeals to them.\
\
The small fee paid by the user is then distributed to the artist's wallet (except the fees required by the network to run the contract calls & return the generated images) - or to a vetted creator foundation in the case of public images.

A user can then opt to download these generated images or mint them as NFTs.

<figure><img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1682476107570/3e5e30f6-f75c-4051-bae5-ee61dbb0c12b.gif?auto=format,compress&#x26;gif-q=60&#x26;format=webm" alt=""><figcaption><p>Images generated on Waterlily.ai</p></figcaption></figure>

## Video

{% embed url="<https://youtu.be/J90uzzfIrAI>" %}
Wes Floyd talks about Waterlily.ai
{% endembed %}

## Find out more about Waterlily.ai

<table data-card-size="large" data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td></td><td><strong>Try out Waterlily.ai for yourself!</strong></td><td></td><td><a href="https://www.waterlily.ai">https://www.waterlily.ai</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-411de7ef9022d889ce5d8eb5994e8a6408e6cdc1%2FWaterlily%20Squarish.png?alt=media">Waterlily Squarish.png</a></td></tr><tr><td></td><td><strong>Contribute to the Github or read the code</strong></td><td></td><td><a href="https://github.com/bacalhau-project/Waterlily">https://github.com/bacalhau-project/Waterlily</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-71ef221484620ab1cfa2f95b09f2083b464b7438%2Foutercore%20waterlily.jpeg?alt=media">outercore waterlily.jpeg</a></td></tr></tbody></table>


# Extra Labs x Lilypad

3D Mapping on Lilypad

This is a work-in-progress proof of concept utilizing Extra Labs and Lilypad to create a Lilypad module for running multiple geospatial jobs.

The goal is to provide the end-to-end adapter for users to operate on the point cloud data to generate reconstructed maps.

Extra Labs is revolutionizing the way maps are built by leveraging collaborative acquisition and deep learning to produce affordable, detailed 3D models of cities and territories.

To achieve this proof of concept, the Lilypad module will integrate with Extra Labs' platform, allowing users to submit their geospatial data, whether collected via drones, aircraft, delivery vehicles, or smartphones. This data is processed through advanced algorithms to create detailed 3D reconstructions.

The decentralized nature of the platform ensures that data providers are compensated fairly for their contributions through a blockchain-based reward system. This approach not only democratizes access to 3D mapping technologies but also ensures continuous and up-to-date data acquisition, enhancing the accuracy and detail of the generated maps.

Users should be able to easily interact with the platform, upload their data, and receive high-quality 3D models in return. This process is designed to make advanced geospatial mapping accessible to a wide range of users, from urban planners and architects to developers and hobbyists.

To find out more, please visit the [Extra Labs website](https://www.extralabs.xyz/).


# DefiKicks (HackFS Winner 2023)

DefiKicks is a decentralized, community-governed Data DAO on Filecoin that democratizes DeFi data aggregation and TVL calculations

{% embed url="<https://stream.mux.com/qyFKcIg2FRHpuy5DQ3d9xgzZFWSe8aXSsEwra3meIpg/high.mp4>" fullWidth="false" %}
DefI Kicks Video
{% endembed %}

{% hint style="info" %}
See the Lilypad Example at [Drand Off-chain Voting](/lilypad/archive/older-versions/lilypad-v0-deprecated/examples/drand-off-chain-voting)
{% endhint %}

DefiKicks is a decentralized, community-governed Data DAO (Decentralized Autonomous Organization) that operates on the Filecoin network, with the goal of democratizing DeFi data aggregation and TVL (Total Value Locked) calculations. It seeks to improve transparency and foster collaboration through a system of on-chain adapter proposals, off-chain voting, and interoperable token rewards.

<div data-full-width="true"><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-0447f6f1048b581499f568ee38c0c8a2bdea027f%2Fdefi-kicks-dashboard.png?alt=media" alt="Defi Kicks Dashboard" width="41%"> <img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-c0149941476a58d5b65fd6a721929cb5bebf70a7%2Fdefi-kicks-propose.png?alt=media" alt="Defi Kicks Propose" width="40%"></div>

At its core, DefiKicks is a counter to the centralization issues prevalent in traditional DeFi data aggregators such as DefiLlama. By shifting the TVL calculations to be performed on-chain, DefiKicks ensures data authenticity and transparency. In essence, it functions similarly to a version control system like 'Git,' allowing anyone to propose a new TVL adapter for a project.

Adapter proposals are initiated on-chain via an IPFS hash, with the adapter code hosted on IPFS. Voting is conducted off-chain using the time-lock encryption of Drand, with results brought back on-chain through Bacalhau Lilypad, a trusted oracle. Upon community approval, adapters are registered on-chain in the 'Registry' contract, which maintains a transparent and immutable record of approved adapters.

Active participants in the voting process who align with the majority are rewarded with KICK tokens. The issuance of these tokens is governed by a smart contract, incentivizing participation and wise decision-making.

Following adapter registration, the decentralized backend begins calculating data points using the code hosted on IPFS. This process employs Lit's Programmable Key Pairs and Lit Actions, with all TVL results hosted on a Ceramic Stream. The results are then viewable by users on the frontend, further promoting an open and transparent DeFi ecosystem.

DefiKicks uses a suite of modern technologies to power its platform, including smart contracts for governance, the LIT protocol for decentralized data governance calculation, Ceramic for decentralized data storage, libp2p for secure peer-to-peer network connections, ApeCoin for alternative voting options, Bacalhau for off-chain jobs execution, Filecoin FEVM for contract interactions, IPFS for decentralized data storage, Web3.Storage for reliable data transfer, and Drand for unbiased off-chain voting.

In essence, DefiKicks, with its combination of technologies, proposes a novel, transparent, and decentralized approach to DeFi data. It is a testament to the possibilities of decentralized governance and data analytics, aiming to usher in a new era of trust and accessibility in DeFi analytics.

## Architecture

![](https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-bb1e478f5143f0631f51a44cb437462d733747b2%2Fimage.png?alt=media)

DefiKicks operates on a foundation of decentralized technologies and protocols to ensure transparent, accessible DeFi data.

It uses smart contracts on Filecoin's FEVM for governance, including adapter proposal, voting, and registration.

The LIT protocol allows for secure, decentralized data calculations, while Ceramic facilitates decentralized data storage.

Lilypad + Bacalhau carries out off-chain jobs such as reward computation and vote tallying.

IPFS and Web3.Storage ensure secure and efficient data storage and transfer, while Drand provides unbiased off-chain voting through time-lock encryption.

## Find out more about DefiKicks.xyz

<table data-view="cards"><thead><tr><th></th><th data-hidden></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td>DefiKicks Live Site</td><td><a href="https://defikicks.xyz/">https://defikicks.xyz/</a></td><td><a href="https://defikicks.on.fleek.co/">https://defikicks.on.fleek.co/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-706b42402d59ad094e3ef5f567f87bb6537d52df%2FScreenshot%202023-06-27%20at%207.58.56%20pm.png?alt=media">Screenshot 2023-06-27 at 7.58.56 pm.png</a></td></tr><tr><td>DefiKicks Github</td><td></td><td><a href="https://github.com/md0x/defikicks">https://github.com/md0x/defikicks</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-71064313abe492e0afce7853cfc4288fcccf5f64%2Fgithub.png?alt=media">github.png</a></td></tr><tr><td>HackFS Submission</td><td></td><td><a href="https://ethglobal.com/showcase/defikicks-b1wo0">https://ethglobal.com/showcase/defikicks-b1wo0</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-263b3e69009ef9e05a672d535425364937cdf850%2FFwMoaTuWIAch8j1.png?alt=media">FwMoaTuWIAch8j1.png</a></td></tr></tbody></table>


# Agents

AI agents on Lilypad

AI agents can autonomously process information, make decisions and interact with users or systems. This section explores different use cases for AI agents on Lilypad, providing examples and step-by-step guides on how to build and run them using the network.

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td>AI Oncologist Agent</td><td>A multi-agent system that searches, extracts, and analyzes oncology research papers for key insights.</td><td><a href="/lilypad/use-cases-agents-and-projects/agents/ai-oncologist-agent">AI Oncologist Agent</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FomKXbKaBG6Asb0kRyngV%2Foutput.png?alt=media&amp;token=f0cf6f79-3806-49f3-8cfe-1614009f4780">output.png</a></td></tr><tr><td>RAG Support Agent</td><td>Retrieves context and generates AI-driven support responses using the Lilypad Network.</td><td><a href="/lilypad/use-cases-agents-and-projects/agents/rag-support-agent">RAG Support Agent</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FMpikjTLfp3jMYBy9dRqu%2Frag.png?alt=media&amp;token=7a21a9e3-f032-46c2-9cd8-de22b5233a48">rag.png</a></td></tr><tr><td>HubSpot-LinkedIn Agent</td><td>An agent-based system that monitors HubSpot for recently modified contacts, fetches enriched data from LinkedIn, and updates HubSpot automatically.</td><td><a href="https://docs.lilypad.tech/lilypad/use-cases-agents-and-projects/agents/hubspot-linkedin-agent">https://docs.lilypad.tech/lilypad/use-cases-agents-and-projects/agents/hubspot-linkedin-agent</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FisfqJL2Pusl9i0imIKQv%2Foutput.png?alt=media&amp;token=8a0cc598-1beb-491b-9408-2edd88b65ef3">output.png</a></td></tr></tbody></table>


# AI Oncologist Agent

A research agent template implemented as an AI oncologist. Read papers, isolate important information, and produce actionable reports.

{% embed url="<https://www.youtube.com/watch?v=5Hq3lUobrN4>" %}
Altruistic AI Agents - Lilypad Oncologist
{% endembed %}

## Agent Details

The AI Oncologist serves as a Research agent template that can be used to conduct a wide range of research. The Agent is an intelligent system for analyzing research papers using a multi-agent approach.

The system consists of three main agents:

#### Paper Relevance Agent

* Searches through PDF documents in the documents directory
* Uses embeddings and cosine similarity for initial filtering
* Verifies relevance using LLM-based analysis
* Returns a list of most relevant paper filenames

#### Top Paragraphs Agent

* Extracts text from identified papers
* Splits content into manageable chunks
* Scores paragraph relevance using LLM
* Returns top-scoring paragraphs with relevance scores

#### Text Query Agent

* Analyzes provided text passages
* Generates focused answers to specific queries
* Uses contextual understanding to provide accurate responses

## Getting Started

A guide to launch the Research agent locally and run inference on the Lilypad Network with the [Lilypad Inference API](https://github.com/Lilypad-Tech/HubSpot-LinkedIn-AgentSync).

### Prerequisites

* Python 3.8+
* OpenAI API key or compatible API (e.g., DeepSeek)
* PDF files

### Installation

1. Clone the repository:

```bash
git clone https://github.com/mavericb/ai-oncologist.git
cd ai-oncologist
```

2. Install required packages:

```bash
pip install -r requirements.txt
```

3. Create a `.env` file in the project root with the following variables:

```env
# OpenAI-Like API configuration
BASE_URL="https://api.deepseek.com"
OPENAI_API_KEY=your_deepseek_api_key
MODEL="deepseek-chat"

ANURA_BASE_URL=https://anura-testnet.lilypad.tech
ANURA_API_KEY=your_anura_api_key
ANURA_MODEL=phi4:14b

# Search configuration
MAX_RESULTS=3
SIMILARITY_THRESHOLD=0.3
```

* `BASE_URL`: API endpoint for OpenAI Compatible LLM service (default: "<https://api.deepseek.com>")
* `OPENAI_API_KEY`: Your OpenAI Compatible API key (for example: [DeepSeek API Docs](https://api-docs.deepseek.com/api/deepseek-api))
* `MODEL`: One OpenAI Compatible model to use (default: "deepseek-chat")
* `ANURA_BASE_URL`: API endpoint for the Anura LLM service (default: "<https://anura-testnet.lilypad.tech>")
* `ANURA_API_KEY`: Your Anura API key (get it here: [Lilypad Inference API Docs](https://docs.lilypad.tech/lilypad/developer-resources/inference-api))
* `ANURA_MODEL`: The Anura model to use (default: "phi4:14b")
* `MAX_RESULTS`: Maximum number of papers to return (default: 3)
* `SIMILARITY_THRESHOLD`: Minimum similarity score for document selection (default: 0.3)

### Usage

1. Place your PDF research papers in the `documents/` directory.
2. Run the main script:

```bash
python AIOncologist.py
```

## Resources

* [GitHub](https://github.com/mavericb/ai-oncologist)


# RAG Support Agent

A retrieval-augmented generation (RAG) agent that retrieves relevant context and generates AI-powered responses using the Lilypad Network.

The Lilypad RAG Support Agent is a Retrieval-Augmented Generation (RAG) AI assistant that retrieves relevant information and generates AI-powered responses using the Lilypad Network. It enables automated support and troubleshooting by leveraging vector search and AI-based text generation.

## **Agent Details**

* **Context-Aware Responses** – Uses `all-MiniLM-L6-v2` embeddings to retrieve relevant information.
* **AI-Powered Answer Generation** – Sends retrieved context and user query to the Lilypad API, which processes it using Llama3 8B.
* **Customizable Knowledge Base** – Modify the agent’s context source (`issues.md`) to adapt it for different use cases.

### **How It Works**

1. **Embedding Queries with `all-MiniLM`**
   * Converts user queries and stored knowledge into dense vector embeddings for similarity-based retrieval
2. **Retrieving Relevant Context**
   * Searches a pre-indexed database to find the most relevant information.
3. **Generating Responses with Llama3 using the** Lilypad API
   * Sends retrieved context and user prompt to the Lilypad API, where Llama3 8B generates a structured response

### **Expanding to Your Own Support Agent**

The Lilypad RAG Support Agent can be adapted to different projects by modifying its retrieval source.

#### **Updating the Knowledge Base (`issues.md`)**

By default, the agent retrieves information from `issues.md`, a markdown file containing troubleshooting steps.

To customize:

* Open `issues.md` in the repository.
* Replace or expand the content with relevant support information for your project.
* Format the content clearly to improve retrieval accuracy.
* Restart the agent to index the updated knowledge base.

For more advanced use cases, the agent can be extended to support multiple files or external knowledge sources.

## **Getting Started**

A guide to launch the RAG support agent locally and run inference on the Lilypad Network with the [Lilypad Inference API](https://github.com/Lilypad-Tech/HubSpot-LinkedIn-AgentSync).

#### **Get a Lilypad API Key**

Sign up at [Anura API](https://anura.lilypad.tech) and generate an API key.

#### **Clone the Repository**

```sh
git clone https://github.com/PBillingsby/lilypad-rag-agent.git
cd lilypad-rag-agent
```

#### **Configure Your API Key**

Export your Lilypad API Token as an environment variable:

```sh
export LILYPAD_API_TOKEN="your-api-key-here"
```

To make it persistent, add it to `~/.bashrc` or `~/.zshrc`.

### **Setup**

Ensure Python 3 is installed, then run:

```sh
pip install -r requirements.txt
```

### Usage

After setting up the API key and dependencies, the agent is ready to process queries using Lilypad’s AI-powered retrieval system.

#### **Run the Agent**

Execute the agent from the project's root directory:

```sh
python3 cli.py
```

### **Resources**

* [Lilypad Docs](https://docs.lilypad.tech)
* [Lilypad Anura API](https://anura.lilypad.tech/)
* [Source code](https://github.com/PBillingsby/lilypad-rag-agent)


# HubSpot-LinkedIn Agent

An agent-based system that monitors HubSpot for recently modified contacts, fetches enriched data from LinkedIn, and updates HubSpot automatically.

## Agent Details

Automate customer relationship management, internal company management, etc. with the [HubSpot-LinkedIn agent](https://github.com/Lilypad-Tech/HubSpot-LinkedIn-AgentSync)! Scale inference using the Lilypad network deploying any of the models supported by the [Lilypad Inference API](https://docs.lilypad.tech/lilypad/developer-resources/inference-api).

This workflow serves as a proof of concept (PoC) illustrating the capabilities of agents to interface with platform APIs, specifically those of HubSpot and RapidAPI.&#x20;

Github repo [here](https://github.com/Lilypad-Tech/HubSpot-LinkedIn-AgentSync).

## Getting Started

A guide to launch the HubSpot-LinkedIn agent locally and run inference on the Lilypad Network with the [Lilypad Inference API](https://github.com/Lilypad-Tech/HubSpot-LinkedIn-AgentSync).

### Prerequisites

* Python 3.8+
* A HubSpot account with a Private App token
  * [Sign up for HubSpot](https://www.hubspot.com/)
  * [Create a Private App](https://developers.hubspot.com/docs/guides/apps/private-apps/overview)\
    *Steps*: Log in > Settings > Integrations > Private Apps > Create private app > Name it > Scopes: `crm.objects.contacts` (read/write contacts) > Create app > Click app name > Auth tab > View access token > Copy token
* A RapidAPI account with a LinkedIn Data API key
  * [Sign up for RapidAPI](https://rapidapi.com/signup)\
    *Steps*: Create an account > Subscribe to [LinkedIn Data API](https://rapidapi.com/rockapis-rockapis-default/api/linkedin-data-api) > Go to [Developer Dashboard > Authorization](https://rapidapi.com/developer/authorization) > Select "default-application" > Copy "Application Key"
* An OpenAI-compatible API
  * [Lilypad Inference API](https://example.com/anura-api-docs) (Sign up and generate an API key)'
  * Choose the model you would like to use

### Setup

1. **Clone the Repository**

   ```
   git clone https://github.com/mavericb/HubSpot-LinkedIn-AgentSync.git
   cd HubSpot-LinkedIn-AgentSync
   ```
2. **Install Dependencies**

   ```
   pip install -r requirements.txt
   ```

   *Note*: If `requirements.txt` is missing, install:

   ```
   pip install python-dotenv hubspot-api-client requests schedule phi
   ```
3. **Configure Environment Variables**
   * Create a `.env` file:

     ```
     touch .env
     ```
   * Add your keys:

     ```
     HUBSPOT_ACCESS_TOKEN=your-private-app-token
     RAPIDAPI_KEY=your-default-app-token
     API_KEY=your-openai-compatible-key
     BASE_URL=https://anura-testnet.lilypad.tech/api/v1  # Adjust for your provider
     AGENT_MODEL=qwen2.5:7b  # Adjust for your provider
     ```

### Usage

1. **Start the Daemon**

   ```
   python main.py
   ```

   * Checks HubSpot every minute for contacts modified in the last 7 days (or since `last_processed.json`).
   * Fetches LinkedIn data and updates HubSpot.
   * Runs for 5 minutes (adjustable in `main.py`).
2. **Monitor Logs**
   * Look for "Found X contacts", "HubSpot update: ...", "Attempting to save last\_processed timestamp".
   * If no contacts are found, `last_processed.json` won’t save—edit a HubSpot contact to test.

### File Structure

* `main.py`: Entry point and daemon logic.
* `/tools/hubspot_tools.py`: HubSpot API utilities (fetch/update contacts).
* `/tools/linkedin_tools.py`: LinkedIn API utility (fetch profile data).
* `/agents/hubspot_agent.py`: Agent for updating HubSpot.
* `/agents/linkedin_agent.py`: Agent for fetching LinkedIn data.


# AI Tooling

AI tooling using Lilypad

AI tooling provides the essential infrastructure and frameworks that support the development, deployment and monitoring of AI-powered applications. This section highlights tools that help streamline workflows, enhance visibility and enable customization when building with AI on Lilypad, with examples and guides for integrating them into your projects.

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td>Anything LLM</td><td>GPT Desktop Client for Lilypad</td><td><a href="/lilypad/tooling-plugins-and-integrations/ai-tooling/anything-llm">Anything LLM</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FhpdsAFWogo4eoi7591dZ%2Fanythingllm.png?alt=media&amp;token=268f74a1-e0d8-4159-9890-f1ed9b3a0ad8">anythingllm.png</a></td></tr><tr><td>Rivet</td><td>Agentic Workflows</td><td><a href="/lilypad/tooling-plugins-and-integrations/ai-tooling/rivet">Rivet</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F2Uux8EtRhpQoulrfYgZS%2Frivet.png?alt=media&amp;token=e8b1b7e7-dca1-4371-83a7-e45eaaa9128e">rivet.png</a></td></tr><tr><td>Lilypad Llama3 Chatbot</td><td>AI-powered chatbot running on the Lilypad Network using Llama3 for intelligent responses.</td><td><a href="/lilypad/tooling-plugins-and-integrations/ai-tooling/lilypad-llama3-chatbot">Lilypad Llama3 Chatbot</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FlASS8vBcVYw5mYDjocVW%2Fchatbot.png?alt=media&amp;token=fb309bc6-f879-4f52-b057-297762b5b136">chatbot.png</a></td></tr><tr><td>Lilypad ML Workbench</td><td>A platform for running, fine-tuning and deploying AI models on the Lilypad Network, with support for Jupyter notebooks and multi-module workflow</td><td><a href="/lilypad/tooling-plugins-and-integrations/ai-tooling/lilypad-ml-workbench">Lilypad ML Workbench</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FLJSF4kDXrMZtN2X0K6Qo%2Fml.png?alt=media&amp;token=2b44d08b-dd25-488d-a3f0-f58b52b9720c">ml.png</a></td></tr></tbody></table>


# Anything LLM

Tutorial: Using Lilypad x AnythingLLM for Desktop LLM Client

{% embed url="<https://www.youtube.com/watch?v=QzRMfm3Tvwk>" %}

## **Getting Started**

### **Prerequisites:**

* A basic understanding of the Lilypad API and Anything LLM
* Access to an Anura Dashboard account

### **Set up Open AI Endpoint in Rivet**

1. Download and Install AnythingLLM  from <https://anythingllm.com/>
2. Start AnythingLLM  App.
3. Navigate to the Settings page by clicking wrench icon at bottom left
4. Click on LLM under AI Providers
5. Select Generic OpenAI under LLM Provider
6. In the BaseURL input field, paste the URL:\
   &#x20; [https://anura-testnet.lilypad.tech/api/v1](https://anura-testnet.lilypad.tech/api/v1/chat/completions)

### **Obtain an Anura API Key**

1. Log in to your Anura account at <https://anura.lilypad.tech/>
2. Click Get Started
3. Create a new API key by clicking the “Create API Key” button.
4. Copy the generated API key and paste it into the APIIey input field on the LLM Screen.

### Specify Chat Model Name

1. Paste in "llama3.1:8b"

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FOP32hVac3SM7v9MKxhqo%2Fimage.png?alt=media&amp;token=4399379d-dd62-481c-a154-d409982f7255" alt=""><figcaption></figcaption></figure>


# Rivet

Tutorial: Using Lilypad x Rivet for Generating Narratives with Anura

{% embed url="<https://www.youtube.com/watch?v=zSeEK-6HZU8>" %}

## **Getting Started**

### **Prerequisites:**

* A basic understanding of the Lilypad API and Rivet
* Access to an Anura Dashboard account
* Basic knowledge of workflow configuration

### **Set up Open AI Endpoint in Rivet**

1. Download and Install Rivet from <https://rivet.ironcladapp.com/>
2. Start Rivet App.
3. Navigate to the Settings page. App>Settings
4. Click on the “Open AI” tab.
5. In the Open AI endpoint input field, paste the URL:\
   &#x20; <https://anura-testnet.lilypad.tech/api/v1/chat/completions>

### &#x20;**Obtain an Anura API Key**

1. Log in to your Anura account at <https://anura.lilypad.tech/>
2. Click Get Started
3. Create a new API key by clicking the “Create API Key” button.
4. Copy the generated API key and paste it into the Open AI API key input field in Rivet.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FDLJh5947FCrJvRKp8iMK%2Fimage.png?alt=media&amp;token=abf4df70-6330-4335-91da-8f07a52d0c20" alt=""><figcaption></figcaption></figure>


# Lilypad Llama3 Chatbot

An AI-powered chatbot that generates responses using the Lilypad API and Llama3.

The **Lilypad Llama3 Chatbot** is a conversational agent designed to deliver engaging, real-time AI-powered interactions. Leveraging the Llama3 8B model through the Lilypad API, this chatbot provides context-aware and dynamic responses, making it useful for applications such as automated support, interactive Q\&A and virtual assistance.

### **Features**

* **AI-Powered Conversations** – Utilizes Llama3 8B via the Lilypad Network to generate intelligent responses.
* **Lilypad API Connectivity** – Simple integration with Lilypad, allowing flexible use of all available models on the Lilypad API.

This chatbot can be extended with:

* **Memory & Context Awareness** – Store conversation history for more personalized interactions.
* **External APIs** – Integrate with knowledge bases, search engines, or database lookups.
* **Multi-Model AI** – Swap Llama3 with other AI models as needed.

### **How It Works**

1. **User Input** – The chatbot interface captures the user's message.
2. **API Request** – The message is sent to the Lilypad API, which runs the job on the Llama3 model.
3. **Response Generation** – Llama3 processes the input, considers context and generates a natural language response.
4. **Response Display** – The response is rendered in the chatbot interface.

### **Prerequisites**

* [**Node.js**](https://nodejs.org/en) – 18 LTS or newer

### **Installation**

1. Get [Anura API key](https://anura.lilypad.tech/).
2. Clone the repository:

   ```sh
   git clone https://github.com/PBillingsby/lilypad-llama3-chatbot.git
   cd lilypad-llama3-chatbot
   ```
3. Install dependencies: `npm install`
4. Add the Lilypad API key to `.env`: `LILYPAD_API_TOKEN=<ANURA_API_KEY>`
5. Start the development server: `npm run dev`

### **Usage**

Access the chatbot at: `http://localhost:3000` and enter prompt to start a conversation.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F7yGKJ2kbgdiNx2uVZyfU%2FScreenshot%202025-03-12%20at%2012.14.17%E2%80%AFPM.png?alt=media&amp;token=4d7563d4-9098-4621-ae47-1ee16c842a2a" alt=""><figcaption><p>Example use</p></figcaption></figure>

### Resources

* [Source code](https://github.com/PBillingsby/lilypad-llama3-chatbot)


# Lilypad ML Workbench

A developer platform for using the Lilypad network to train and run AI models

The [Lilypad ML workbench](https://github.com/arsen3d/LilypadWorkbench) provides ML researchers, businesses, and more a simple interface for leveraging the Lilypad network.

The ML workbench provides an interface to:

* Run models that are currently available on the Lilypad GPU network
* Add a new model to run on the network as a [Lilypad module](https://docs.lilypad.tech/lilypad/developer-resources/build-a-job-module)
* Leverage Jupyter notebooks with Lilypad
* Create multi-module flows
* Fine tune models
* Create an AI agent

In order to run an AI model on the Lilypad network, a docker image must be created for the program using the [Lilypad module spec](https://docs.lilypad.tech/lilypad/developer-resources/build-a-job-module).

{% hint style="info" %}
The ML workbench product is currently in beta testing and is not yet provided as a hosted solution. Reach out in the Lilypad [Discord](https://discord.gg/zSErHwZSyC) for more info!
{% endhint %}

## AI Inference

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FvOmoax1cRre1nANIGHq9%2FMLWorkbench1.png?alt=media&amp;token=7b8375b6-a692-4a6c-a580-37a61c4453b3" alt="" width="563"><figcaption></figcaption></figure>

Run existing Lilypad modules with a simple text or image prompt. The workbench will output an IPFS CID with the result.

Implement the Lilypad module allowlist and cache the models locally before they are needed! Pin the files to a local IPFS node packaged with the workbench platform.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FHhQiHrN5VVlrcVxigInQ%2FMLWorkbench5.png?alt=media&amp;token=beab6b31-7a83-484b-a6ce-c4629dce64eb" alt=""><figcaption></figcaption></figure>

## Finetuning

Import a dataset and use the the "Data" tool to finetune a model for a specific end user. Create a Lilypad module and save the module on IPFS. More on this soon!

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F4krivniK1v8Alm9jdE1J%2FMLWorkbench9.png?alt=media&amp;token=cf44975f-10b7-46d8-be4f-fdea36a47909" alt=""><figcaption></figcaption></figure>

## Create an AI agent and run on Lilypad

The ML workbench provides a simple interface for creating and training AI agents. The Lilypad team is currently testing a variety of models and fine tuning techniques to determine the optimal path forward for this feature. More on this soon!

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F7caaI6W9fAfhpNLjVP9v%2FMLWorkbench10.png?alt=media&amp;token=34ee846c-cd5b-4dfe-bc4b-c95117f272b8" alt=""><figcaption></figcaption></figure>

### Resources

* [Source code](https://github.com/arsen3d/LilypadWorkbench)


# Integrations

Built with Lilypad

This section features hands-on tools and resources designed to support developers building with Lilypad. From code editor extensions to social app integrations, it highlights ways to prototype, automate and ship AI jobs more efficiently. Explore how these utilities fit into real workflows and help extend the power of the network.

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td>VS Code Helper Extension</td><td>Enhances development with smart suggestions, automation and simple Lilypad integration.</td><td><a href="/lilypad/tooling-plugins-and-integrations/integrations/vs-code-helper-extension">VS Code Helper Extension</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FsKa8DwSvXRsw8MDE9AH6%2Fvscode.png?alt=media&amp;token=ec08210f-0207-4145-b309-d3a1cdc9945e">vscode.png</a></td></tr><tr><td>Farcaster Frame</td><td>Allow users to generate AI-powered outputs directly within interactive Farcaster posts</td><td><a href="/lilypad/tooling-plugins-and-integrations/integrations/farcaster-frame">Farcaster frame</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F87tMOMUOQHU3h8sKzhT3%2Fframe.png?alt=media&amp;token=9b081432-3238-4a90-969e-080f3d234cdd">frame.png</a></td></tr></tbody></table>

Need some inspiration? Visit the [Helpful Resources](/lilypad/developer-resources/useful-links) page!


# VS Code Helper Extension

An integrated development environment for running AI workloads on the Lilypad Network directly within Visual Studio Code.

The **Lilypad VS Code Extension** allows developers to interact with the Lilypad Anura API within Visual Studio Code. By highlighting code blocks, users can query the available LLMs on Lilypad  for explanations, improvements, or suggestions, receiving responses in a formatted webview panel.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FRkrbUBRqt5KujxSuFITV%2Fanura-vscode-gif.gif?alt=media&amp;token=e52e9344-381c-4382-aaa3-4eabca72c634" alt=""><figcaption><p>VS Code extension in action</p></figcaption></figure>

### Features

* **AI-Powered Code Assistance** – Select any code in your editor, choose an AI model, and ask Lilypad questions about it
* **Formatted Responses** – View AI-generated insights in a structured webview panel
* **Secure API Configuration** – Store your Lilypad API token securely
* **Context Menu Integration** – Quickly access the extension’s features via the right-click menu

### Installation

From the Repository

1. Get [Anura API key](https://anura.lilypad.tech/).
2. Clone the repo: `git clone git@github.com:PBillingsby/lilypad-vscode-extension.git`
3. Install the extension in VS Code: `code --install-extension lilypad-vscode-extension.vsix`
4. Add Anura API key to `.env`:  `LILYPAD_API_TOKEN=<ANURA_API_KEY>`
5. If you make changes to the extension and need to recompile it:
   1. Recompile the code: `npm run compile`
   2. Repackage the extension: `vsce package`
   3. Reinstall the updated `.vsix` file: `code --install-extension lilypad-vscode-extension.vsix`

### **Usage**

1. **Select a block of code** in your editor.
2. **Right-click** and select **"Ask Lilypad about this code"**, or:
   * Open the **Command Palette** (`Ctrl+Shift+P`) and select **"Ask Lilypad about this code"**.
3. **Choose an AI model** to process the query.
4. **Enter your question** related to the selected code.
5. **Wait for Lilypad AI** to process your query.
6. **View the AI’s response** in the webview panel that opens.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F2hcGs8sm6UG1bJ3N4Fhi%2FScreenshot%202025-03-10%20at%203.05.20%E2%80%AFPM.png?alt=media&amp;token=3f447908-5ff5-4644-88ee-3190012c7ae9" alt=""><figcaption><p>Example use</p></figcaption></figure>

### Resources

* [Source code](https://github.com/PBillingsby/lilypad-vscode-extension)


# Farcaster frame

A guide for running Lilypad jobs in a Farcaster frame

In this guide we will be building a Farcaster frame for local development that runs a Lilysay prompt on the Lilypad Network.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fxt6jZMieaN6YXFoKHatc%2Finitial.gif?alt=media&amp;token=523aee82-2773-4ecb-be9c-df1fa0b8d48c" alt=""><figcaption></figcaption></figure>

Users of the frame can input a prompt and a generate an image. The generated image will appear in the box above the input and will allow a user to view the ASCII art as an image.

{% hint style="info" %}
This is an example for running a local version as inspiration for developers. The Lilypad CLI will need to be wrapped and the project hosted as an API to run in production and be used in a [Farcaster Frame](https://docs.farcaster.xyz/developers/frames/).
{% endhint %}

## Project Setup

### Generate project

For this frame, create a NextJS 14 app:

```bash
npx create-next-app lilypad-farcaster-example
cd lilypad-farcaster-example
```

Next, install the required dependencies. [Coinbase onchain kit](https://github.com/coinbase/onchainkit) is used in this project:

```bash
npm install @coinbase/onchainkit uuid sharp@0.32.6 path fs
```

Open the project in your code editor.

### Get testnet tokens

We will need to fund a wallet with Lilypad and Arbitrum Sepolia testnet tokens. Follow the first 2 sections labelled "Setting up MetaMask" and "Funding your wallet" from our [Quick Start](/lilypad/quickstart/cli) docs.

### Add `.env.local`

Add the following into your `.env.local` file. The private key is used to run the CLI jobs. Make sure that your `.env.local` file is added to the `.gitignore` file as your private key should not be exposed or pushed.

{% hint style="info" %}
Your `WEB3_PRIVATE_KEY` can be retrieved from the MetaMask account details menu. For more info, check out the [official guide from MetaMask](https://support.metamask.io/managing-my-wallet/secret-recovery-phrase-and-private-keys/how-to-export-an-accounts-private-key/) on how to get a your private key. Please do not push your private key to GitHub.
{% endhint %}

1. `WEB3_PRIVATE_KEY=YOUR_PRIVATE_KEY`
2. `NEXT_PUBLIC_BASE_URL=http://localhost:3000`

### Install Lilypad CLI

Run through this guide to [install the Lilypad CLI on your machine](/lilypad/quickstart/cli/installation). Select the CLI User script.

### Run Framegear

Framegear is a simple tool provided by the `@coinbase/onchainkit` package that allows you to run and test your frames locally without publishing the frame.

{% hint style="info" %}
We will be using Framegear for this build. However, there are other libraries that can be used for this project instead of Framegear.
{% endhint %}

In a separate terminal, clone down the `onchainkit` repo and run `Framegear`:

```tsx
git clone https://github.com/coinbase/onchainkit.git
cd onchainkit/framegear
npm i
npm run dev
```

Navigate to [`http://localhost:1337`](http://localhost:1337/) and keep that window open for when we start to write the frame.

## Configuration for frame

We will need to set up the metadata for our Next.js application that includes Farcaster frame information. It will configure the elements and a URL for frame posting, while also specifying Open Graph metadata for improved social sharing.

In `app/page.tsx`, add the following before the `Home` function declaration:

```tsx
import { getFrameMetadata } from '@coinbase/onchainkit/frame';
import type { Metadata } from 'next';
const frameMetadata = getFrameMetadata({
  buttons: [
    {
      label: "Let's go!"
    }
  ],
  image: `${process.env.NEXT_PUBLIC_BASE_URL}/initial.gif`,
  post_url: `${process.env.NEXT_PUBLIC_BASE_URL}/api/frame?id=1`,
});

export const metadata: Metadata = {
  title: 'Lilysay Farcaster frame',
  description: 'Initial frame',
  openGraph: {
    title: 'Lilysay Farcaster frame',
    description: 'Initial frame',
    images: ["/initial.png"],
  },
  other: {
    ...frameMetadata,
  },
};
```

## UI Components

The UI elements for this frame are all rendered in the `app/api/route.ts` file, which acts as a request handler for different routes or endpoints within the web application. It defines the logic for handling user input, generating responses, and managing application state. The main functions include processing user prompts, handling status checks, and generating images asynchronously.

### Routes

Here’s how the routes are structured for this frame:

* `/api/frame?action=input`: This route displays the initial user interface, which includes a text input field for the prompt and a button to submit the form. The user interface also updates dynamically based on the processing status, such as showing a placeholder image or the final generated image.
* `/api/frame?action=submit`: This route processes the user input. When a prompt is submitted, the server initiates the image generation process asynchronously. While the image is being generated, the user sees a loading state, and they can check the progress.
* `/api/frame?action=check`: This route checks the status of the image generation process. It updates the frame with either a completed image, an error message if the generation fails, or the processing state if the image is still being generated.
* We also include a fallback just in case an error occurs during the processing of job.
* `/api/frame?action=save`: Though not explicitly included, this could be an additional route for handling the logic of saving the generated image to a location for future access.

### `generateImage`

The `generateImage` function handles the user input and generates the final image for display by utilizing the functions in `app/services/cli.ts`. It ensures that the image is generated asynchronously and the result is available for display in the frame, or handled properly in case of any errors during the generation process.

### Frame images

Throughout the interaction process, various images are used. These images serve as visual cues during each step of the frame, such as when the user is prompted for input, while the image is being processed and once the final result is ready or if an error occurs.

To obtain the images used in this guide, save them from [the source code here](https://github.com/PBillingsby/lilypad-frame-example/tree/main/public).

## Create `route.ts`

Create the `api/frame` directories in your project. Inside of the `app/api/frame` directory, create a file named `route.ts` and add the following code:

```tsx
import { NextRequest, NextResponse } from 'next/server';
import { v4 as uuidv4 } from 'uuid';
import { runCliCommand } from '@/app/services/cli';
import * as fs from 'fs/promises';
import path from 'path';

interface RequestData {
  prompt: string;
  status: 'processing' | 'completed' | 'error';
}

const requestStore: { [key: string]: RequestData } = {};

async function getResponse(req: NextRequest): Promise<NextResponse> {
  const searchParams = req.nextUrl.searchParams;
  const action = searchParams.get("action") || "input";
  const requestId = searchParams.get("id");
  const data = await req.json();
  const prompt = data.untrustedData?.inputText;

  switch (action) {
    case "input":
      return new NextResponse(`<!DOCTYPE html><html><head>
        <title>Input Prompt</title>
        <meta property="fc:frame" content="vNext" />
        <meta property="og:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/enter-prompt.png" />
        <meta property="fc:frame:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/enter-prompt.png" />
        <meta property="fc:frame:button:1" content="Submit" />
        <meta property="fc:frame:input:text" content="Enter your prompt" />
        <meta property="fc:frame:post_url" content="${process.env.NEXT_PUBLIC_BASE_URL}/api/frame?action=submit" />
      </head></html>`);

    case "submit":
      if (prompt) {
        const id = uuidv4();
        requestStore[id] = { prompt, status: 'processing' };

        // Start the image generation process asynchronously
        generateImage(id, prompt);

        // Immediately return the loading state
        return new NextResponse(`<!DOCTYPE html><html><head>
          <title>Request Submitted</title>
          <meta property="fc:frame" content="vNext" />
          <meta property="og:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/loading.gif" />
          <meta property="fc:frame:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/loading.gif" />
          <meta property="fc:frame:button:1" content="Check Status" />
          <meta property="fc:frame:post_url" content="${process.env.NEXT_PUBLIC_BASE_URL}/api/frame?action=check&id=${id}" />
        </head></html>`);
      }
      break;

    case "check":
      if (requestId && requestId in requestStore) {
        const request = requestStore[requestId];
        switch (request.status) {
          case 'completed':
            const imageUrl = `${process.env.NEXT_PUBLIC_BASE_URL}/result.png`;
            return new NextResponse(`<!DOCTYPE html><html><head>
              <title>Result</title>
              <meta property="fc:frame" content="vNext" />
              <meta property="og:image" content="${imageUrl}" />
              <meta property="fc:frame:image" content="${imageUrl}" />
              <meta property="fc:frame:button:1" content="New Request" />
              <meta property="fc:frame:post_url" content="${process.env.NEXT_PUBLIC_BASE_URL}/api/frame?action=input" />
            </head></html>`);
          case 'error':
            return new NextResponse(`<!DOCTYPE html><html><head>
              <title>Error</title>
              <meta property="fc:frame" content="vNext" />
              <meta property="og:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/error.png" />
              <meta property="fc:frame:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/error.png" />
              <meta property="fc:frame:button:1" content="Try Again" />
              <meta property="fc:frame:post_url" content="${process.env.NEXT_PUBLIC_BASE_URL}/api/frame?action=input" />
            </head></html>`);
          case 'processing':
            return new NextResponse(`<!DOCTYPE html><html><head>
              <title>Processing</title>
              <meta property="fc:frame" content="vNext" />
              <meta property="og:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/loading.gif" />
              <meta property="fc:frame:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/loading.gif" />
              <meta property="fc:frame:button:1" content="Check Again" />
              <meta property="fc:frame:post_url" content="${process.env.NEXT_PUBLIC_BASE_URL}/api/frame?action=check&id=${requestId}" />
            </head></html>`);
        }
      }
      break;
  }

  // Fallback response
  return new NextResponse(`<!DOCTYPE html><html><head>
    <title>Error</title>
    <meta property="fc:frame" content="vNext" />
    <meta property="og:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/error.png" />
    <meta property="fc:frame:image" content="${process.env.NEXT_PUBLIC_BASE_URL}/error.png" />
    <meta property="fc:frame:button:1" content="Start Over" />
    <meta property="fc:frame:post_url" content="${process.env.NEXT_PUBLIC_BASE_URL}/api/frame?action=input" />
  </head></html>`);
}

// Runs the CLI command and saves the result as an image
async function generateImage(id: string, prompt: string) {
  try {
    const imageBuffer: any = await runCliCommand(prompt);
    const imagePath = path.join(process.cwd(), 'public', 'result.png');
    await fs.writeFile(imagePath, imageBuffer);
    requestStore[id] = { prompt, status: 'completed' };
  } catch (error) {
    console.error('Error generating image:', error);
    requestStore[id] = { prompt, status: 'error' };
  }
}

export async function POST(req: NextRequest): Promise<Response> {
  return getResponse(req);
}

export const dynamic = 'force-dynamic';
```

## Create `cli.ts`

This is where the execution of the Lilysay job happens. It will run the job, wait for it to finish and then create an image from the return value.

Inside of the `app` directory, create a new directory named `services` and inside of that create a file named `cli.ts`. The functions inside this file will allow us to send a prompt to the Lilypad Network using the user prompt and a predefined command that runs asynchronously in the terminal. Once the command is executed, Lilypad processes the input through its Lilysay module and outputs the results in the form of an ASCII image, which is then converted into a displayable image using an SVG-to-PNG transformation.

Here are the 3 functions inside this file:

* **`createImageBufferFromAscii`**: Converts ASCII text into an SVG image and then uses the `sharp` library to convert the SVG into a PNG image buffer. This allows the display or saving of an image representation of the ASCII text.
* **`runCliCommand`**: Executes a Lilypad CLI command to process the user's input text, captures the command's output, and converts it into an image buffer. It handles the entire process of running the command, capturing the output, and managing errors.
* **`extractStdoutFilePath`**: Parses the CLI command's stdout to extract the file path where the Lilypad CLI has saved the output. It uses a regex pattern to identify the path in the command's output.

The following code snippet demonstrates the process:

```typescript
import { spawn } from 'child_process';
import { promises as fs } from 'fs';
import * as path from 'path';
import { fileURLToPath } from 'url';
import sharp from 'sharp';

const __filename = fileURLToPath(import.meta.url);
const __dirname = path.dirname(__filename);

const MODULE_VERSION = "cowsay:v0.0.4";

// Function to generate an image buffer from ASCII text
async function createImageBufferFromAscii(asciiText, aspectRatio = 1.91) {
  const fontSize = 14;
  const lineHeight = fontSize + 6;
  const padding = 20;

  const lines = asciiText.split('\\n');

  const textHeight = lines.length * lineHeight + padding * 2;

  let width, height;
  if (aspectRatio === 1.91) {
    width = Math.max(textHeight * aspectRatio, 800);
    height = textHeight;
  } else if (aspectRatio === 1) {
    width = height = Math.max(textHeight, 800);
  }

  const escapeXML = (unsafe) => {
    return unsafe.replace(/&/g, "&amp;")
      .replace(/</g, "&lt;")
      .replace(/>/g, "&gt;")
      .replace(/"/g, "&quot;")
      .replace(/'/g, "&#039;");
  };

  const escapedLines = lines.map(line => escapeXML(line));

  const textWidth = Math.max(...escapedLines.map(line => line.length)) * fontSize * 0.6;
  const xPosition = Math.max((width - textWidth) / 2, 10);

  // Generate SVG markup with the ASCII content
  const svgImage = `
    <svg width="${width}" height="${height}" xmlns="<http://www.w3.org/2000/svg>">
      <rect width="100%" height="100%" fill="white" />
      <style>
        .text-content { font-family: monospace; font-size: ${fontSize}px; fill: black; white-space: pre; }
      </style>
      <text x="${xPosition}" y="${padding}" class="text-content">
        ${escapedLines.map((line, index) => `<tspan x="${xPosition}" dy="${index === 0 ? '0' : lineHeight}">${line}</tspan>`).join('')}
      </text>
    </svg>
  `;

  // Convert the SVG to a PNG image buffer using sharp
  return sharp(Buffer.from(svgImage)).png().toBuffer();
}

// Function to run the Lilypad CLI command
export async function runCliCommand(inputs) {
  console.log("Lilypad Starting...");

  const web3PrivateKey = process.env.WEB3_PRIVATE_KEY;
  if (!web3PrivateKey) {
    throw new Error('WEB3_PRIVATE_KEY is not set in the environment variables.');
  }

  // Construct the command to run Lilypad with the user input
  const command = `lilypad run ${MODULE_VERSION} -i Message="${inputs}"`;
  console.log("Command to be executed:", command);

  // Execute the command as a shell process
  return new Promise((resolve, reject) => {
    const child = spawn('bash', ['-c', `export WEB3_PRIVATE_KEY=${web3PrivateKey} && ${command}`]);

    let stdoutData = '';
    let stderrData = '';

    // Capture stdout from the CLI command
    child.stdout.on('data', (data) => {
      stdoutData += data.toString();
      console.log(`Stdout: ${data}`);
    });

    child.stderr.on('data', (data) => {
      stderrData += data.toString();
      console.error(`Stderr: ${data}`);
    });

    child.on('close', async (code) => {
      if (code !== 0) {
        reject(new Error(`Process exited with code ${code}`));
        return;
      }

      if (stderrData) {
        reject(new Error(stderrData));
        return;
      }

      console.log("Process completed successfully!");

      try {
        // Extracts the file path, reads the ASCII content and converts it to an image buffer
        const stdoutFilePath = extractStdoutFilePath(stdoutData);
        const asciiContent = await fs.readFile(stdoutFilePath, 'utf-8');
        const imageBuffer = await createImageBufferFromAscii(asciiContent);
        resolve(imageBuffer);
      } catch (error) {
        reject(new Error(`Error processing output: ${error.message}`));
      }
    });

    child.on('error', (error) => {
      reject(new Error(`Error with spawning process: ${error.message}`));
    });
  });
}

// Helper function to extract the stdout file path from the CLI output
function extractStdoutFilePath(stdoutData) {
  const match = stdoutData.match(/cat (\/tmp\/lilypad\/data\/downloaded-files\/\w+\/stdout)/);
  if (!match || !match[1]) {
    throw new Error('Stdout file path not found in CLI output');
  }
  return match[1];
}
```

* **Sending the Request**: The user's input text is passed directly into the **Lilypad CLI command** using a shell process. The input text is embedded within the command's arguments and executed asynchronously in the terminal.
* **Handling the Response**: After the CLI command completes, the output is captured and processed. The response includes the file path to the generated ASCII image, which is then read from the file system and converted into a PNG image for further use.
* **Error Handling**: If an error occurs during the execution of the CLI command or file processing, it is logged to the console, and the process is terminated with appropriate error messaging.

## Testing

The Framegear server is running. Next, run your local server in your frame project. We will need to make sure it is running on port 3000:

```bash
npm run dev
```

Navigate to the Framegear host [http://localhost:1337](http://localhost:1337/) and you will see an input labeled "Enter your frame URL". Add <http://localhost:3000> to that and click "Fetch".

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fm4nxU4IjYGM0uRbwTNdA%2Fframegear.png?alt=media&amp;token=b7fda38f-7344-4732-a27c-3668f8e04dab" alt=""><figcaption></figcaption></figure>

You should now see your frame and be able to interact with it. Enter a prompt to display in the Lilysay ACSII image!&#x20;

## Rendering results

As the job is processed, a “Check Status” button will be displayed. Clicking this will check if the job has been completed. Once a job is completed, the results will be displayed in the frame. If there is an issue along the way, an error message will be displayed and the user will be prompted to try again.

## Potential Use Cases

Running Lilysay jobs is just one of the ways you can utilize Lilypad in a frame, but you can run jobs with any available modules on the Lilypad Network. Some of these include:

* Stable Diffusion image
* Stable Diffusion video
* Ollama

## Resources

* [Source code](https://github.com/PBillingsby/lilypad-frame-example)
* [Farcaster frames documentation](https://docs.farcaster.xyz/developers/frames/)


# Lilypad Workflows

A set of n8n workflows using Lilypad AI Models

At Lilypad, we’ve integrated n8n with the Lilypad API to automate dynamic workflows, combining human input, AI generation and multi-platform actions.

Using Lilypad’s OpenAI-compatible endpoints inside n8n lets you:

* 🧠 Source and enrich content: Pull from databases like Notion, Airtable, or Google Sheets and enhance it with AI.
* 🎨 Generate custom outputs: Create AI-written text, summaries, or even modify images dynamically.
* 🔄 Automate publishing: Push results to platforms like Twitter, Discord, Slack, or your own apps.
* 📈 Track and update: Monitor workflow status and feed results back into your systems.

Everything runs automatically inside n8n, with Lilypad API endpoints providing powerful, cost-free AI capabilities at the heart of it.

Imagine building workflows that review customer feedback with AI, generate personalized emails, summarize research papers, trigger alerts from on-chain events, or even create full content campaigns, all without writing a single script. With n8n and Lilypad, AI-driven automation is at your fingertips.

Coming soon:

<table data-view="cards"><thead><tr><th></th></tr></thead><tbody><tr><td>Marketing Workflows</td></tr><tr><td>Life workflows</td></tr></tbody></table>


# Hardware Requirements

This page overviews the hardware requirements to operate a Lilypad Network node. It's important to note that these requirements continuously evolve as the network grows. If you have questions or suggestions, please join our [Discord](https://discord.gg/tnE8SMmsxW) or open a pull request on the [Lilypad documentation repo](https://github.com/Lilypad-Tech/lilypad-docs).&#x20;

Minimum Hardware Requirements

* **Processor**: Quad-core x64 Processor or better
* **RAM**: 32GB (see additional details below)
* **Internet**: Internet: 250Mbps download, 100Mbps upload (minimum)
* **GPU**: NVIDIA GPU with a minimum of 8GB VRAM (see additional details below)
* **Storage:** SSD with at least 500gb of free space

## GPU Requirements

Each model operating on Lilypad has specific VRAM (Video Random Access Memory) requirements directly related to the model's complexity and computational demands. For running a Lilypad Resource Provider (RP) with multiple GPUs, a guide using Proxmox can be found [here](https://github.com/Lilypad-Tech/lilypad-tools/blob/main/proxmox/multi-gpu-proxmox.md).

* **Base Requirement:** The simplest model on Lilypad requires a GPU with at least 8GB of VRAM. This is the minimum required to participate in computational tasks on the Lilypad network.

### Model-Specific VRAM Requirements

The capability of your GPU to manage multiple or more complex Lilypad jobs is enhanced by the amount of VRAM available:

* **Standard Models (SDXL,** **Ollama):** Require at least 8GB of VRAM.
* **Advanced Models:** Require 14GB+ of VRAM.

### Hardware Compatibility

* GPUs with 8GB of VRAM are limited to running models like SDXL, which fit within this specification. Larger GPUs with higher VRAM are required for more demanding models like SDV, which needs at least 14GB of VRAM.

#### For example:

* A node with a GPU containing 8GB of VRAM can execute Lilypad [SDXL](https://github.com/Lilypad-Tech/lilypad-module-sdxl-pipeline) module jobs, which require a minimum of 8GB of VRAM.
* Larger capacity GPUs are needed for heavier compute models like [SDV](https://github.com/Lilypad-Tech/lilypad-module-sdv-pipeline), which require at least 14GB of VRAM.

## RAM Requirements

Lilypad uses the Resource Provider's GPU to load models, initially requiring the temporary storage of the data in the system's RAM. In a production environment with RP Nodes, it is important to have enough RAM to support the model and the underlying system's operational processes.

### Minimum RAM Requirements:

* **Base Requirement:** A minimum of 16GB of RAM is required, with at least 8GB dedicated to the model and another 8GB allocated for system operations and Lilypad processes.

## Additional Considerations

* **Wallets for each GPU:** You need a separate account for each GPU you want to set up on the network. The wallet you use for your account must have both ETH (to run smart contracts on Ethereum) and Lilypad (LP) tokens in order to receive funds for jobs) on the network.

<figure><img src="https://github.com/Lilypad-Tech/lilypad-docs/blob/main/.gitbook/assets/eth-lp-wallet.png.png" alt="" width="355"><figcaption><p>"a wallet funded with both ETH and LP tokens"</p></figcaption></figure>

* **Larger Models:** Jobs involving more substantial models will demand additional RAM. It's important to note that adequate VRAM alone is insufficient; your system must also have enough RAM to load the model into the GPU successfully. Without sufficient system RAM, the model cannot be loaded into the GPU, regardless of available VRAM.

## Apply to our closed Beta Resource Provider program!

{% embed url="<https://forms.gle/XaE3rRuXVLxTnZto7>" %}


# Run a node

Instructions for setting up a Resource Provider (node) on the public Lilypad testnet using Docker, including obtaining necessary funds and installing required software.

## Prerequisites

{% hint style="info" %}
Lilypad RPs currently only support Linux installs. Running a RP on Windows is currently experimental.
{% endhint %}

* Linux (*Ubuntu 22.04* LTS)
* Nvidia GPU
* [Nvidia drivers](https://ubuntu.com/server/docs/nvidia-drivers-installation)
* [Docker](https://docs.docker.com/engine/install/ubuntu/) (Ubuntu install)
* [Nvidia Docker drivers](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html) (NVIDIA Container Toolkit)

{% hint style="info" %}
For a more in-depth look at the requirements to run a Lilypad node, please refer to the [hardware requirements](/lilypad/resource-providers/hardware-requirements) documentation.
{% endhint %}

## Network information and Testnet tokens

The testnet has a base currency of ETH, as well as a utility token called LP. Both are used for running nodes. To add a node to the testnet, follow these steps:

### Metamask Configuration

We recommend using MetaMask with custom settings to make things easier. Once you have it installed and setup, here are the settings you need to use:

{% hint style="info" %}
Network name: Arbitrum Sepolia

New RPC URL: <https://sepolia-rollup.arbitrum.io/rpc>

Chain ID: 421614

Currency symbol: ETH

Block explorer URL: (leave blank)
{% endhint %}

For a step by step guide on adding the network, please refer to our [Setting up MetaMask documentation](/lilypad/quickstart/cli/setting-up-your-wallet).

### Fund your wallet with ETH and LP

To obtain testnet LP, use the [Lilypad faucet](http://faucet.lilypad.tech) and enter your ETH address.

To obtain testnet ETH, use a third party [Arbitrum Sepolia testnet faucet](https://arbitrum.faucet.dev/ArbSepolia) and enter your ETH address.

{% hint style="info" %}
If you need more tokens and already have Sepolia ETH, use the [official Arbitrum bridge](https://bridge.arbitrum.io/) to transfer tokens over to Arbitrum Sepolia.
{% endhint %}

The faucet will give you both ETH (to pay for gas) and LP (to stake and pay for jobs).

## Setup Arbitrum RPC (Optional) <a href="#setup-arbitrum-rpc" id="setup-arbitrum-rpc"></a>

The Lilypad Network uses the Arbitrum Sepolia Testnet to settle compute transactions. When a transaction is ready to be saved on-chain, Lilypad cycles through a list of public Arbitrum Sepolia RPC endpoints using the endpoint that settles first to save the compute transaction.

Resource Providers have the option to [setup their own Arbitrum RPC endpoint](https://docs.lilypad.tech/lilypad/hardware-providers/setup-arbitrum-rpc) using Alchemy instead of using the default public RPC endpoints.

A personal RPC endpoint helps RPs to avoid reliability issues with the public RPC endpoints used by Lilypad ensuring rewards can be earned and jobs can be run consistently. RPs running a personal RPC endpoint contribute to the fault tolerance and decentralization of the Lilypad Network! Read more in the Alchemy Arbitrum [docs](https://docs.alchemy.com/reference/arbitrum-api-quickstart).

## Docker Compose Setup

Before we start the Rp with the Docker setup, retrieve the private key from the wallet set up earlier in this guide. For guidance on exporting your private key, refer to [this official MetaMask guide](https://support.metamask.io/managing-my-wallet/secret-recovery-phrase-and-private-keys/how-to-export-an-accounts-private-key/). Once the private key has been retrieved, proceed to initialize the Docker containers using the commands provided below.

### 1. Export Your Private Key

{% hint style="info" %}
The same WEB3\_PRIVATE\_KEY cannot be used for both a RP and using the Lilypad CLI. If a WEB3\_PRIVATE\_KEY has already been used to run jobs with the CLI, make a new one and [fund the wallet](/lilypad/quickstart/cli/setting-up-your-wallet). You can \`unset $WEB3\_PRIVATE\_KEY\` if you want to use a different one to run your RP.
{% endhint %}

Before starting, export your private key from MetaMask. Follow the official MetaMask guide for instructions on safely exporting your private key.

### 2. Download the Docker Compose Configuration

Use `curl` to download the `docker-compose.yml` file from the Lilypad GitHub repository.

```bash
LATEST_VERSION=$(curl -s https://api.github.com/repos/Lilypad-Tech/lilypad/releases/latest | sed -n 's/.*"tag_name": "\(.*\)".*/\1/p')

curl -o docker-compose.yml "https://raw.githubusercontent.com/Lilypad-Tech/lilypad/$LATEST_VERSION/docker/docker-compose.yml"
```

### 3. Check for Existing Containers

{% hint style="warning" %}
If any containers named `resource-provider`, `ipfs`, or `watchtower` are already in use, they will need to be stopped before running this setup to avoid naming conflicts.
{% endhint %}

You can check if these containers are running with:

```bash
docker ps -a | grep -E "resource-provider|ipfs|watchtower"
```

If they are running, stop them with:

```
docker stop <container_name>
```

If there are still conflicts when trying to running with the docker-compose file, remove the containers:

```shell-session
docker rm <container_name>
```

{% hint style="info" %}
Before moving to the next step, ensure older versions of lilypad and bacalhau [are not running](/lilypad/resource-providers/troubleshooting) as systemd services.
{% endhint %}

### 4. Start the Resource Provider

Start the Lilypad containers using Docker Compose:

```bash
WEB3_PRIVATE_KEY=<your_private_key> docker compose up -d
```

To include a custom RPC URL:&#x20;

```
WEB3_PRIVATE_KEY=<your_private_key> WEB3_RPC_URL=wss://arb-sepolia.g.alchemy.com/v2/your-alchemy-id docker compose up -d
```

{% hint style="warning" %}
You must not reuse your compute node key as a client, even for testing: this will result in failed jobs and will negatively impact your compute node since the wallet address is how nodes are identified on the network.
{% endhint %}

### 5. Monitor Your Node

Use the following command to check the status of the resource provider and bacalhau.

```bash
docker logs resource-provider
```

```bash
docker logs bacalhau
```

Use the following command to view the containers running after starting Docker Compose.

```bash
docker ps
```

NOTE: Occasional websocket errors in your RP logs are normal and nothing to worry about. These are simply the result of standard internet fluctuations.

Our solver may also restart periodically, which can temporarily interrupt your connection.

Rest assured that these brief disconnections will not affect your rewards in any way.

## Update Lilypad version for Docker RP

When a new version of Lilypad is released, it is critical for Resource Providers to update their installations to ensure compatibility and ability to run Lilypad jobs.&#x20;

### 1. If a Docker RP is running stop the system (if the node is not running, disregard this first step)

```bash
docker compose down
```

### 2. View all Docker images

```bash
docker images
```

### 3. Delete old Docker images that are duplicates for Lilypad (Bacalhau, Lilypad)

```bash
docker rmi <image_name_or_id> 
```

### 4. Follow Docker RP install instructions

Using the Lilypad Docker [RP install instructions](/lilypad/resource-providers/run-a-node) setup a new RP and run.

## View Lilybit\_ rewards

To view your Lilybit\_ rewards, visit one of the following dashboards and paste your node's public address into the input:

* [Grafana dashboard](https://grafana.lilypad.tech/d/adxhou3o1q8sga/rewards-per-wallets?orgId=1\&refresh=1m)

## Troubleshooting

[Here](https://docs.lilypad.tech/lilypad/resource-providers/troubleshooting) are some common troubleshooting techniques when it comes to your resource provider using Docker.


# Update node

Instructions to update the Docker Lilypad Resource Provider (RP)

To update a Lilypad RP, remove any previous versions of Lilypad on the instance and then follow the instructions to setup a Docker RP.

* If the Lilypad install for Linux was previously used on the RP, first [remove](https://docs.lilypad.tech/lilypad/resource-providers/update-node#ensure-all-processes-from-linux-install-are-stopped-and-removed)[ the install](https://docs.lilypad.tech/lilypad/resource-providers/update-node#ensure-all-processes-from-linux-install-are-stopped-and-removed). Then follow the [install instructions](https://docs.lilypad.tech/lilypad/resource-providers/update-node#install-and-run-a-new-version-of-lilypad).
* If the RP previously used a Docker Lilypad version,[  previous Docker containers and images](https://docs.lilypad.tech/lilypad/resource-providers/update-node#update-lilypad-version-for-docker-rp), then follow the [install instructions](https://docs.lilypad.tech/lilypad/resource-providers/update-node#install-and-run-a-new-version-of-lilypad).

## Ensure all processes from Linux install are stopped and removed

*This only applies if you had a Linux version installed*

### 1. Stop all systemd processess

```bash
sudo systemctl stop bacalhau
```

```bash
sudo systemctl stop lilypad-resource-provider
```

### 2. Disable the systemd services:

```bash
sudo systemctl disable bacalhau
```

```bash
sudo systemctl disable lilypad-resource-provider
```

### 3. Delete the service files from the systemd directory.

{% hint style="info" %}
*Note: Be extremely careful when using `sudo` to remove files.*
{% endhint %}

```bash
sudo rm /etc/systemd/system/bacalhau.service
```

```bash
sudo rm /etc/systemd/system/lilypad-resource-provider.service
```

### 4. Reload the systemd daemon to apply the changes.

```bash
sudo systemctl daemon-reload
```

## Remove old Docker containers and images <a href="#update-lilypad-version-for-docker-rp" id="update-lilypad-version-for-docker-rp"></a>

### 1. If a Docker RP is running stop the system (if the node is not running, disregard this first step)

```bash
docker compose down
```

### 2. You can check the status of the containers with:

```bash
docker ps -a | grep -E "resource-provider|ipfs|watchtower"
```

If they are running, stop them with:

```bash
docker stop <container_name>
```

&#x20;Remove the containers:

```bash
docker rm <container_name>
```

### 3. View all Docker images

```bash
docker images
```

### 4. Take note of the IMAGE ID for each lilypad image (resource-provider, bacalhau, and watchtower)&#x20;

Delete old Docker images that are duplicates for Lilypad (Bacalhau, Lilypad)

```bash
docker rmi <IMAGE ID>  
```

{% hint style="info" %}
&#x20;*`docker volume prune` can also be used to remove specific images that aren't being used.*
{% endhint %}

## Install and Run a new Version of Lilypad

{% hint style="info" %}
If a RP was running on the Lilypad Testnet in 2024, it is recommended to create a new wallet when joining the network again for the RP Beta program. If a RP wants to use the same wallet, feel free to try running the RP and let our team know if any issues with running jobs are experienced.

If a wallet has been used to run CLI jobs on Lilypad, this wallet cannot be used for a RP). See the Docs for [Setting up a Metamask Wallet](/lilypad/quickstart/cli/setting-up-your-wallet)
{% endhint %}

### 1. Export WEB3\_PRIVATE\_KEY as an environment variable

```bash
export WEB3_PRIVATE_KEY=<your-private-key>
```

### 2. Use `curl` to download the `docker-compose.yml` file from the Lilypad GitHub repository.

```
LATEST_VERSION=$(curl -s https://api.github.com/repos/Lilypad-Tech/lilypad/releases/latest | sed -n 's/.*"tag_name": "\(.*\)".*/\1/p')

curl -o docker-compose.yml "https://raw.githubusercontent.com/Lilypad-Tech/lilypad/$LATEST_VERSION/docker/docker-compose.yml"
```

### 3. Start your Lilypad Node

```bash
WEB3_PRIVATE_KEY=<your_private_key> docker compose up -d
```

OR start with your own RPC URL

```bash
WEB3_PRIVATE_KEY=<your_private_key> WEB3_RPC_URL=wss://arb-sepolia.g.alchemy.com/v2/your-alchemy-id docker compose up -d
```

## Monitor the Resource Provider <a href="#id-5.-monitor-your-node" id="id-5.-monitor-your-node"></a>

Use the following command to check the status of the resource provider and bacalhau.

```bash
docker logs resource-provider
```

```bash
docker logs bacalhau
```

Use the following command to view the containers running after starting Docker Compose.

```bash
docker ps
```

A healthy, updated node should have all containers started, a preflight check, and be adding a resource offer.<br>

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FsOYX2c4h7vyUPx6Xz6TB%2FScreenshot%20from%202025-02-19%2009-40-43.png?alt=media&amp;token=7c5f3f7c-feb7-4ee6-a771-9cf12a1d5004" alt=""><figcaption></figcaption></figure>


# Troubleshooting

Common FAQs when running a Lilypad node

{% hint style="info" %}
Please view these resources below for FAQs!

* [Lilybit Rewards info](https://blog.lilypadnetwork.org/update-to-the-lilybit-rewards-calculation)
* [Resource Provider Points dashboard](https://rp-points.lilypad.tech/)
* [Lilybit Leaderboard](https://info.lilypad.tech/leaderboard) (under maintenance)
  {% endhint %}

## Don't see your issue below?

For **complex issues, bug reports, or feature requests,** open a discussion in the Lilypad-Tech Github organization discussion [board](https://github.com/orgs/Lilypad-Tech/discussions).

* Navigate to the discussion [board](https://github.com/orgs/Lilypad-Tech/discussions), select "New Discussion", choose "rp-issues", and fill out the template.
* Without a discussion opened, our team will not be able to support the problem.

**For quick questions or minor issues,** use the Lilypad Discord [#i-need-help](https://discord.com/channels/1212897693450641498/1230231823674642513) channel and provide the following info.

* Description (including Lilypad version running on your node)
* Hardware Info (including Linux/Windows version)
* Related blockchain/ETH addresses of transaction hashes
* Output Logs - `sudo systemctl status lilypad-resource-provider`
* Related links/urls
* Screenshots

{% hint style="info" %}
**IMPORTANT:** When sharing screenshots of your logs or node information, make sure to **remove or block out any references to your node's private key.** Keeping your private key safe and away from the public eye is crucial!
{% endhint %}

## FAQ

### Common issues to check first!

1. Ensure the RP (node) is running the latest [Lilypad version](https://docs.lilypad.tech/lilypad/resource-providers/run-a-node#id-2.-download-the-docker-compose-configuration) for your preferred environment
2. Does the RP have [enough](/lilypad/resource-providers/run-a-node#fund-your-wallet-with-eth-and-lp) Lilypad Tokens (LP) and Arbitrum ETH?
3. Updating and restarting the Lilypad services regularly (daily) is encouraged throughout IncentiveNet.

## Run a node (Resource Provider - RP)

### How do I keep track of Lilypad version releases and other important announcements?

The [updates-rp](https://discord.com/channels/1212897693450641498/1256179769356189707) Discord channel is the primary location for Resource Provider announcements. Announcements in this channel are also posted on the Lilypad [updates](https://updates.lilypad.tech/) page.

Lilypad supports Resource Providers using the Docker find instructions [here](/lilypad/resource-providers/run-a-node).&#x20;

If this doesn't solve the problem, [raise a ticket](https://discord.com/channels/1212897693450641498/1230231823674642513) with our team.

### **How can I check the status of my Lilypad RP once it's running?**

To check if the RP is running use the following command:

```bash
docker logs resource-provider
```

This will give a live output from the Lilypad node. The logs will show the node running and accepting jobs on the network.

Run the following command to get info from Bacalhau

```bash
docker logs bacalhau
```

### Can I become a Lilypad RP with just a CPU and no GPU?

Resource Providers can run on the Lilypad Network without a GPU, however only hardware with a GPU is currently rewarded with Lilybit\_ rewards.

### **Checking Docker Runtime**

To verify your Docker runtime configuration: `sudo docker info | grep Runtimes`

You should see the NVIDIA runtime listed. If you only see: `Runtimes: io.containerd.runc.v2 runc` you will need to configure the NVIDIA runtime.

### **Configuring NVIDIA Runtime**

If the NVIDIA runtime is not showing up or you're experiencing issues, try the following:

1\. Configure the NVIDIA Container Toolkit runtime: `sudo nvidia-ctk runtime configure --runtime=docker --set-as-default`\
2\. Restart the Docker service: `sudo systemctl restart docker`

### **Overview of Docker setup**

For a comprehensive overview of your Docker setup, use: `docker info`. This command provides detailed information about your Docker daemon configuration.

### My docker RP is turning on, but showing errors stating it is not providing a Resource Offer.

Typically this occurs when an old version of Lilypad is still running on the instance.&#x20;

1. Ensure the Bacalhau and Lilypad systemd services are stopped and removed.

```bash
sudo systemctl stop bacalhau
```

```bash
sudo systemctl stop lilypad-resource-provider
```

2. Disable the systemd services so they start on boot:

```bash
sudo systemctl disable bacalhau
```

```bash
sudo systemctl disable lilypad-resource-provider
```

3. Delete the service files from the systemd directory.

\*\*Note: Be extremely careful when using `sudo` to remove files.

```bash
 sudo rm /etc/systemd/system/bacalhau.service
```

```bash
 sudo rm /etc/systemd/system/lilypad-resource-provider.service
```

4. Reload the systemd daemon to apply the changes.

```bash
sudo systemctl daemon-reload
```

### Can I run a Lilypad RP on Windows?

Lilypad RPs currently only support Linux installs. Running a RP on Windows is currently experimental.

### How do I run multiple GPU’s on one server?

Recommendation guide using Proxmox found [here](https://github.com/Lilypad-Tech/lilypad-tools/blob/main/proxmox/multi-gpu-proxmox.md). More on this coming soon!

### Can I run multiple Lilypad RPs on one GPU?

No, this would be considered detrimental to the Lilypad network and cannot be rewarded at this time.

In order to run jobs successfully, Lilypad RPs must ensure all resources on the server are available. If a single GPU is running many RPs and splitting the server resources between them, this causes each RP to not have enough resources to run a Lilypad job.

### How do I setup a personal RPC for Arbitrum Sepolia?

Here's a quick guide on setting up your own RPC for a Lilypad node.

{% embed url="<https://www.youtube.com/watch?v=INKpdrfF0Xs>" %}

### CompatNotSupportedOnDevice Error

The CUDA version of the RP does not match the GPU driver. Please refer to this [Nvidia GPU Driver guide](https://docs.nvidia.com/deploy/cuda-compatibility) to repair installation.

### Unknown Error Code: 222

Indicates that the CUDA version of the RP is incorrect. Install the CUDA version which suitable for the gpu type and compile Lilypad by themselves.

## Lilypad IncentiveNet details

When do Lilybit\_ rewards earned or rewards slashed appear in the Leaderboard? \*\*The leaderboard is currenlty under maintanence, find Lilybits earned by [RPs](https://rp-points.lilypad.tech/) here.

### Are **there required updates needed to maintain my node software with Lilypad on IncentiveNet?**

Resource providers are expected to have the latest Lilypad version installed on their systems. The installation instructions can be found here:

* [Docker](https://docs.lilypad.tech/lilypad/hardware-providers/run-a-node/docker#enable-automatic-updates)

To stay up to date with the latest releases, check the [updates-rp channel](https://discord.com/channels/1212897693450641498/1256179769356189707) in the Lilypad Discord or visit the [Lilypad GitHub](https://github.com/Lilypad-Tech/lilypad). Along the top menu, click the "Watch" dropdown and you will see a section named "Custom". Selecting "Releases" will allow you to get notified of any new releases!

### Can Lilybit rewards be sent to another wallet? Will this function be added in the future?

Currently, it's not possible. However, it's a very good feature request and the team is evaluating!

### I’m getting an error of “invalid hex character 'r' in private key”

This is more than likely due to you trying to export your mnemonic seed phrase instead of the private key. A private key typically appears like this: `4c0883a69102937d6231471b5dbb6204fe512961708279df95b4a2200eb6b5e7` and consists of 64 hexadecimal characters.

Check out the [MetaMask official guide to retrieve your private key](https://support.metamask.io/managing-my-wallet/secret-recovery-phrase-and-private-keys/how-to-export-an-accounts-private-key/).

### How do I setup my Metamask wallet?

View the [docs here](/lilypad/quickstart/cli/setting-up-your-wallet).

### How do I get testnet LP and ETH (Arbitrum Sepolia ETH)?

* [Fund](https://docs.lilypad.tech/lilypad/lilypad-testnet/quick-start/funding-your-wallet-from-faucet) a wallet with LP and ETH
* No ETH or LP in your wallet? ([import custom network and import the tokens](https://docs.lilypad.tech/lilypad/lilypad-testnet/quick-start/setting-up-metamask#import-the-testnet-lp-token))

Join our [Discord](https://lilypad.team/discord) for more help!


# RP-Beta Program

Become an early RP on the Lilypad Network and power the future of Open Access AI

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fimur6QyvqboOcFT4RzAp%2FResource-Providers--1--1.png?alt=media&amp;token=0951fcdf-65dc-4f95-88e4-f4cc24fec3a7" alt=""><figcaption><p>Empowering AI-driven Innovation</p></figcaption></figure>

\
Read more about contributing hardware and earning rewards in Lilypads ongoing RP-Beta Program.

<https://blog.lilypadnetwork.org/resource-provider-beta-program><br>

Become a Resource Provider: [Apply here!](https://docs.google.com/forms/d/e/1FAIpQLSeF7xIHuCpwY0X44dqnl4u3weuvmtd5MkZKY0IPlGck4kHx3w/viewform)


# Lilypad Litepaper

Lilypad: A Protocol for Open Intelligence Infrastructure

***

### Building the Infrastructure for an Open AI Economy

#### Abstract

Lilypad is a coordination protocol for intelligence infrastructure. It provides permissionless access to verifiable compute, model monetisation by default, and cryptographically auditable execution for any AI workload. Lilypad abstracts compute primitives into a modular, trust-minimised protocol that can serve inference, agents, and pipeline-based workflows across open and enterprise systems.

This paper outlines the core thesis, architectural structure, crypto-economic system, and forward-looking roadmap for Lilypad. Our goal: to create a programmable substrate that supports open AI as a public good and enables the next era of agentic computation, composable intelligence, and developer-owned ecosystems.

### 1. Introduction: Infrastructure Shapes Intelligence... and Competitive Edge

#### 1.1 The Technological Convergence Making DeAI Possible

From Turing’s foundational thought experiments to the cloud boom of the 2000s, compute has mirrored, and magnified, our evolving social architectures. Each leap in processing capability has triggered new markets, new models of coordination, and new vulnerabilities. Now, with AI becoming not just an application layer but a decision substrate, the question isn’t just what we build, but where, and who gets to participate.

We’re at a convergence point of three arcs:

* The **Arc of Compute:** from mainframes to hyperscaler cloud to edge devices, with increasing decentralisation and specialisation at the hardware layer.
* The **Arc of AI:** from statistical tooling to foundational models to agentic systems running autonomous workflows.
* The **Arc of Coordination:** from institutionally brokered trust to cryptographic primitives, on-chain settlement, and composable economic logic.

The convergence of open-source AI and crypto-infrastructure is the foundation for a new economic paradigm. Lilypad exists at the intersection of these arcs merging crypto’s programmable market infrastructure with AI’s intelligence primitives to build an AI innovation economy at scale.

#### 1.2 Proprietary AI & Centralised Infrastructure Challenges&#x20;

Today’s centralised AI infrastructure imposes monopolistic control & extractive defaults:

* Compute access intermediated by monopolistic brokers
* Models abstracted behind opaque APIs with limited visibility&#x20;
* Profit & regulatory arbitrage is prioritised over innovation
* Extractive economic value capture leaves foundational contributors unpaid with no native attribution, provenance, or monetisation available

Centralised AI is also struggling to find moat:

* Oligopolies have (so far) failed to lock in regulatory capture
* In-demand AI talent is increasingly ethically aligned with open source AI (eg. Yann LeCun, Mira Murati)
* Proprietary model companies are now competing on distribution channels rather than model quality or capability due to a convergence of factors including&#x20;
  * The Open Source model movement push narrowing the gap of capability between proprietary & open source AI alongside (even Huggingface's $4.5B valuation is driven by network effect)
  * Models, particularly LLMs, approaching inherent limitations due to data saturation and diminishing returns from scaling, suggesting performance improvements for general models are limited
  * The China-US AI race providing new training methods with less resources and competitive open source models

Additionally, centralised infrastructure is struggling to cope with demand and remain cutting edge with

* New data-centres out of date before the GPU cards arrive or not configured for the latest AI needs
* Overhead costs not competitive with more agile alternatives
* Energy & environmental constraints limiting scaling

#### 1.3 The Decentralised AI Advantage

Decentralised AI Ethical Advantage: Community Led & Owned

* AI as a *commons:* owned and governed by the communities that contribute to and use it
* Promotes ethical pluralism, avoiding centralised control and opaque value systems from large tech firms
* Open participation ensures ethical alignment and reduces corporate capture and bias - with local knowledge, underrepresented languages, indigenous ontologies, and non-Western value systems—creating a richer, more inclusive model corpus
* Empowers value-aligned innovation across domains accelerating breakthroughs including BioML, climate, regenerative tech, and rare disease research

Decentralised AI Product Advantage: Open Access & Long-Tail Innovation

* Open participation unlocks a massive long tail of use cases, from hobbyist tools to enterprise-grade models
* Lowers barriers to entry for solopreneurs and domain experts to contribute, run, and monetize specialized models
* Diversity from open participation & access expands the long-tail of use cases fostering a pluralistic AI landscape that better reflects and serves humanity at large

Decentralised AI Technological Advantage: Modular, Scalable, Verifiable Infrastructure

* DePIN&#x20;
  * provides dynamically scalable compute infrastructure with locational advantages
  * provides more efficient compute networks that capture idle compute power
* Market & Monetisation Mechanics
  * Marketplaces provide cost, distribution & performance benefits&#x20;
  * Monetisation pushes state-of-the-art hardware & model accessibility&#x20;
  * Monetisation provides pathways to collective data collection & collaborative training efforts
  * Marketplaces & Monetisation solves energy needs efficiently - driving costs to underlying physics needs.&#x20;
* Trust & Coordination
  * Permissionless coordination capabilities provide broad-based participation and drive collaboration
  * Provenance pipelines & codified trust enable solutions to near-term AI challenges such as data sharing & collection, IP rights and proof of truth and personhood
  * Built-in provenance, module validation, and job reproducibility ensures verifiability and scientific integrity
  * Permissionless coordination supports agent workflows, registry systems, and autonomous pipelines

#### 1.4 DeAI Challenges

Decentralised AI is not without it's challenges:

* Supply-First DePIN Imbalance
  * Many DePIN projects are supply heavy - with web3 being uniquely good at large scale network coordination, hardware for compute and storage is readily available.&#x20;
  * However, many DePIN projects lack a corresponding "demand engine" - with low usage of a fundamentally valuable asset.&#x20;
* Fragmentation & Lack of Collaboration
  * The deAI ecosystem is fragmented: agents, compute networks, model hubs & data platforms are building in silos.
  * A unified coordination layer or interoperable standards are missing, limiting network effects and cross-project capabilities.
  * A collaborative effort across the deAI space is required to compete with centralised solutions.
* Product Immaturity or Pain Points vs Centralised Competitors
  * UX for decentralised compute, model monetisation, and orchestration often lags behind centralised tools
  * Lack of plug-and-play SDKs, agent orchestration tools, or composable pipelines limits adoption outside crypto-native devs.
  * Blockchain payments & primitives can hamper adoption
* Incentive Misalignment Across Layers
  * DePIN operators, model builders, and app developers are incentivised separately: yet success relies on end-to-end alignment.
  * Lack of shared revenue or reputation mechanisms limits module reuse, collaboration, and scale.

While no single project is capable of solving this alone, Lilypad is one of the few projects orchestrating compute, data provenance, and economic incentives in a composable stack capable of collaborative integrations.\
\
Lilypad helps brings the DeAI vision to life by providing a flywheel for DePIN, a distribution path for model training networks, integrations to agentic frameworks, data & storage and a full stack AI service platform for developers, businesses and solopreneurs, all powered by transparent and user-owned economics.

#### 1.5 Lilypad’s Core Philosophy

> "AI should be a public good - not a corporate asset."

Lilypad offers a counter-architecture to centralised models: modular, verifiable & permissionless. It builds a trustless execution layer where models can be exposed, reused, and paid on-chain and reroutes economic gravity back towards creators and contributors.

It’s a programmable substrate for intelligence that expands access, reinforces agency, and creates new terrain for open coordination, utilising the competitive edge of decentralised AI for the collective benefit of humanity.

### 2. Design Philosophy

Lilypad’s architecture is informed by a decade of experiments in decentralized compute: Golem, Truebit, Modicum, Bacalhau, BOINC, and other pioneering systems from academia and the open-source world. Each advanced the state of the art. But none offered the combination of open participation, verifiable execution, and built-in economic coordination.

Golem struggled with adoption and required determinism at the cost of generality. Modicum introduced a promising recomputation-based system, but was constrained by deterministic job assumptions and verifier logic entanglement.\
\
Bacalhau, which underpins our execution layer, was built to solve decentralised job distribution but left incentive mechanisms out of scope. We designed and implemented those incentives as first-class protocol logic: token-driven payments, staking, routing, and verification coordination, all executed through on-chain contracts and off-chain job resolution. Lilypad is an intelligent, incentive-aligned coordination protocol designed to operate compute executions at network scale.

Lilypad is built to enable modular and independent verification, with clear separation from the execution layer. This architectural choice lets the protocol flexibly support deterministic models, probabilistic audits, and log-based validation without compromising performance or developer freedom. Rather than treating auditability as a constraint, Lilypad treats it as a programmable layer for innovation where multiple strategies can coexist and evolve.

We’ve taken lessons from these systems to architect Lilypad differently:

* We decouple verification from execution, enabling multiple verification strategies depending on job type, latency, and value.
* We don’t assume determinism. Instead, we build modularity into verification layers with upgade capability.
* We avoid hardwiring compute to consensus. Lilypad coordinates compute jobs via constraint-solving and smart contract validation, not consensus-layer integration.

Most importantly, we do not assume a monolithic trust architecture. We treat verifiability, coordination, and monetization as primitives.

Lilypad is structured around three core design principles:

1. **Modularity**: Infrastructure should be upgradeable. Each layer of the protocol—from compute to registry to settlement—is designed to evolve independently.
2. **Verifiability**: Every job run should leave a legible, auditable trail. Not all trust needs to be cryptographic—but all trust should be accountable.
3. **Economic symmetry**: Value created by the network should circulate within it. Contributors—whether model authors, solvers, or compute providers—are paid at the protocol level.

### 3. Protocol Architecture

```
┌──────────────────────────────────────────────┐
│              Lilypad Protocol Stack          │
├──────────────────────────────────────────────┤
│         Halo Identity & Auth Layer           │
├──────────────────────────────────────────────┤
│       Ophir Coordination (Smart Contracts)   │
├──────────────────────────────────────────────┤
│          Ampli Model Registry (IP, Metadata) │
├──────────────────────────────────────────────┤
│             Anura Inference API              │
├──────────────────────────────────────────────┤
│         Atlas Compute (Bacalhau / WASM)      │
└──────────────────────────────────────────────┘
```

**Atlas Compute**: Container-based execution using Bacalhau, backed by verifiable job manifests and node-level attestation. Supports WASM and GPU workloads.

**Ampli Registry**: Version-controlled registry for models, modules, and pipelines. Each asset includes signatures, pricing, and lineage.

**Ophir Coordination**: EVM-based contract layer that handles job creation, staking, slashing, and payments. Runs on Arbitrum with upcoming extensions.

**Halo Identity**: Credentialing and authorship framework for model creators, node operators, and agents. Optional enterprise modes and delegated authority.

**Anura API**: Interface layer for querying models, exposing endpoints, and integrating agentic workflows. Compatible with LangChain, OpenRouter, and n8n.

#### Blockchain-based Value

Blockchain is the foundation of Lilypad's coordination logic. It enables atomic settlement, actor-level accountability, and programmable infrastructure at a scale traditional systems cannot replicate.

Smart contracts enforce job flows. Tokens encode trust and incentives directly into the network. No middlemen. No billing overhead. No platform dependencies.

Blockchain offers core properties that general-purpose infrastructure still can't:

* **Instant settlement:** No billing cycles. Compute providers, agents, and creators are paid atomically when jobs are completed.
* **Resilience:** Workload routing happens across a global mesh. There's no single point of failure.
* **Verifiability**: Every job, output, and payment is recorded on-chain, with signatures and manifests.
* **Provenance:** Model authorship and job reuse are recorded cryptographically in the registry.
* **Coordination**: Smart contracts execute logic that would otherwise require middleware or manual enforcement.
* **Distribution**: Anyone with a GPU or a model can participate—from solo developers to research labs to DAOs.

Lilypad flips the script on extractive AI platforms by embedding value flows directly into the infrastructure layer. Model creators, compute providers, mediators, and agents all interact through transparent, programmable incentives. This creates new contribution pathways and aligns every participant to the health and growth of the network. These aren’t speculative mechanisms—they’re live and accruing value on every job.

### 4. Actors and Execution Flow

Lilypad uses a modular actor model, with responsibilities and incentives structured across roles:

| Role               | Function                                             |
| ------------------ | ---------------------------------------------------- |
| Job Creators       | Submit jobs, define inputs, fund usage upfront       |
| Resource Providers | Execute jobs on local or cloud GPUs, earn per output |
| Module Providers   | Publish reusable pipelines and tools                 |
| Solvers            | Match jobs to compute via constraint solving         |
| Mediators          | Sample and verify job execution correctness          |
| Smart Contracts    | Route payments, validate signatures, update state    |

Jobs follow a lifecycle: submission → matching → execution → optional mediation → payment. All state is auditable, and all components are modular.

### 5. Verification and Provenance

Lilypad supports layered verification through signed job manifests, output log digests, and mediation sampling. These are implemented through on-chain job state commitments and optional mediator-sampled audits. Model fingerprinting and FHE are future considerations, but not currently active components of the system.

Read our paper on verification here: <https://arxiv.org/abs/2501.05374>

### 6. Economics and Coordination

Lilypad introduces LILY: a fixed-supply token used to route incentives and economic coordination across the protocol. The system is designed to fairly compensate compute providers, model contributors, and protocol participants, while maintaining integrity through collateral and verifiable actions.

#### 6.1 Payment Modes

* **Pay-as-you-go Execution**: JCs fund jobs at submission. If the job is not run, the unused funds are returned automatically. No prepayment subscriptions, no lock-in.
* **Stake-Weighted Routing**: RPs and Solvers are prioritized by stake and historical performance.

#### 6.2 Actor Value and Rewards

Each actor in the Lilypad ecosystem is assigned a dynamic **value function** Vᵢ(t) based on their activity, contributions, and reputation. This value function determines their share of protocol rewards for a given epoch.

Example: Resource Provider Value

V\_RPᵢ(t) = \[ R̄ᵢ^α₁ \* ( Σ Fⱼ(t) )^α₂ \* C̄ᵢ^α₃ ]

Where:

* R̄ᵢ: average reputation of RP over time window W
* C̄ᵢ: average collateral
* Fⱼ(t): fee earned from job j in timestep t
* α₁, α₂, α₃: weight coefficients

Rewards are distributed proportionally:

Rᵢₖ(t) = \[ Vᵢₖ(t) / ΣVᵢₖ(t) ] \* Rₖ(t)

Where:

* Rₖ(t): total reward pool for actor class k in epoch t
* Vᵢₖ(t): value of actor i in class k

#### 6.3 Collateral Mechanics

To accept jobs, RPs must stake a base collateral amount C\_min, which scales with the job price and reputation. For mediation-eligible jobs:

C\_active ≥ ρ × P\_job, where ρ ≥ 2.5

Higher staked collateral improves job routing probability and early unlock of vesting rewards. Failed jobs or dishonest behavior result in slashing.

This system balances:

* Game Theory Mechanics for compute providers
* Protections for job creators
* Economic throughput that aligns incentives across the network

#### 6.4 Reputation & Value

Each actor has a dynamic reputation score and value metric, used to:

* Filter job preferences
* Prioritize matching
* Allocate protocol rewards

These metrics evolve over time, encouraging consistency and reliability.

### 7. Roadmap

The Lilypad roadmap is structured around three major phases—each aligned to protocol capabilities and the evolving demands of decentralised AI.

#### Phase 1: Compute & Coordination (2024)

| Capability                  | Description                                                                                |
| --------------------------- | ------------------------------------------------------------------------------------------ |
| Job execution pipeline      | Streaming and multi-stage inference workflows with DAG-based job coordination.             |
| Smart contract coordination | Secure, gas-efficient execution payments, state transitions, and refund logic.             |
| Identity and actor roles    | Lightweight on-chain tracking for job creators, solvers, mediators, and compute providers. |
| Validator-in-the-loop       | Optional mediation for job integrity using dispute resolution logic.                       |

#### Phase 2: AI Tooling & Incentives (2025)

| Capability                  | Description                                                                                   |
| --------------------------- | --------------------------------------------------------------------------------------------- |
| SaaS-tier integration       | Anura APIs, hosted endpoints, and fiat billing interface for developers and non-crypto users. |
| Reputation-weighted routing | Actor scores influence job distribution and protocol incentives.                              |
| Persistent job graphs       | Long-running, stateful pipelines for agentic applications.                                    |
| TrainPad                    | Lightweight, distributed finetuning engine with reward-weighted checkpoints.                  |
| Mainnet Deployment & TGE    | Finalisation of token mechanics and mainnet readiness.                                        |

#### Phase 3: Interoperability & Governance (2026)

| Capability               | Description                                                                                           |
| ------------------------ | ----------------------------------------------------------------------------------------------------- |
| ForgeNet                 | Network of remixable, interdependent model forks with royalty logic.                                  |
| Cross-chain execution    | Job layer integration with Taisu and Somnia ecosystems.                                               |
| Governance               | Lilypad DAO activation with protocol parameter control, actor slashing appeals, and grant allocation. |
| Multi-chain pipelines    | Modular jobs distributed across chains, with universal attribution and payment routing.               |
| Additional token utility | Further ecosystem mechanics, including potential L2 expansion or sovereign chain launch.              |

This roadmap ensures Lilypad evolves from job execution into a full composability and coordination layer for decentralised AI systems

### 8. Lilypad SaaS and Enterprise

Lilypad's modular infrastructure powers more than just on-chain coordination. It also supports standalone SaaS deployments and tailored enterprise integrations.

#### 8.1 Lilypad SaaS

For builders who want plug-and-play access to Lilypad's capabilities, our SaaS offering includes:

* Anura-based APIs for running inference on open models
* Hosted dashboards for tracking job execution and usage
* Integration modules for n8n, LangChain, and agent frameworks
* Standard billing interfaces for fiat and non-crypto workflows

SaaS users interact with the full power of the Lilypad job mesh—without needing to touch tokens, wallets, or on-chain logic. It's ideal for developers, startups, and researchers seeking scale and flexibility without overhead.

#### 8.2 Lilypad Enterprise

Enterprise partners can deploy Lilypad in fully private or hybrid environments:

* Dedicated job queues with custom SLAs
* On-prem inference and agent execution
* Compliance tooling for pharma, science, and defence workflows
* Integration with corporate identity and data privacy systems

Enterprise stacks inherit the full Lilypad coordination layer—compute orchestration, verification, monetisation—while retaining local control and data isolation.

Both offerings are backed by the same protocol primitives: smart contracts, verifiable execution, and economic coordination.

### 9. Lilypad: The Coordination Engine for Collective Intelligence at Scale

Lilypad is a programmable coordination system for decentralised intelligence infrastructure.

Imagine swarms of agents coordinating scientific research or building decentralised AI systems - not tied to one API, but operating across a permissionless network with embedded attribution and payment.

Blockchain makes this coordination possible. Payments happen in real time. Coordination logic is encoded into smart contracts. And the system is globally resilient and auditable by default.

Why blockchain matters here:

* **Instant settlement**: No billing cycles. Compute, coordination, and creators are paid atomically.
* **Resilience**: Jobs route across a global mesh, not a single provider.
* **Verifiability**: Cryptographic job manifests, signed results, on-chain records.
* **Provenance**: Model authorship and reuse tracked natively in the registry.
* **Coordination**: Smart contract logic replaces the need for gatekeepers.
* **Distribution**: Actors anywhere can participate—from solo devs to academic labs.

We believe the intelligence economy should be:

* **Verifiable**: Every output traceable, reproducible, auditable.
* **Composable**: Jobs, modules, agents—all interoperable.
* **Fair**: Value flows to contributors, not just platforms.

So we built it. Lilypad is the compute substrate and coordination economy for decentralised AI - a programmable foundation for open-access AI and innovation unlocks.

***

Docs → [https://docs.lilypad.tech](https://docs.lilypad.tech/)\
GitHub → <https://github.com/Lilypad-Tech>

***

**Lilypad.tech | docs.lilypad.tech | github.com/Lilypad-Tech**


# Validation of GPU Compute

Validation of GPU Computation in Decentralized, Trustless Networks

## Overview

The Lilypad Research team published "[Validation of GPU Computation in Decentralized, Trustless Networks](https://arxiv.org/abs/2501.05374)" diving into the complexities of validating successful compute jobs on GPUs. The investigation then explores verification methods and solutions that Lilypad can implement across the network.

Verifying computational processes in decentralized networks poses a fundamental challenge, particularly for Graphics Processing Unit (GPU) computations. The Lilypad Research team conducted an investigation revealing significant limitations in existing approaches: exact recomputation fails due to computational non-determinism across GPU nodes, Trusted Execution Environments (TEEs) require specialized hardware, and Fully Homomorphic Encryption (FHE) faces prohibitive computational costs.&#x20;

To address these challenges, this report explores three verification methodologies adapted from adjacent technical domains: model fingerprinting techniques, semantic similarity analysis, and GPU profiling. Through systematic exploration of these approaches, we develop novel probabilistic verification frameworks, including a binary reference model with trusted node verification and a ternary consensus framework that eliminates trust requirements.&#x20;

These methodologies establish a foundation for ensuring computational integrity across untrusted networks while addressing the inherent challenges of non-deterministic execution in GPU-accelerated workloads.


# V1 Research Documents

These documents provide a background for ongoing Lilypad research.

{% hint style="info" %}
If you have questions or find something in here interesting, please feel free to raise a discussion in the [GitHub](https://github.com/Lilypad-Tech/lilypad-docs) or in [the Lilypad Discord server](https://lilypad.team/discord)!
{% endhint %}

These documents provide a background for the ongoing research on Lilypad. They are primarily focused on the game theory and cryptoeconomics of the protocol, and include an introduction to verifiable computing, an overview of the specific problems we are tackling, a brief overview of prior work, a description of our approach, and mechanisms that we plan to test in simulation.

### Learn more about:

<table data-view="cards"><thead><tr><th data-card-target data-type="content-ref"></th><th></th></tr></thead><tbody><tr><td><a href="/lilypad/research-and-vision/v1-research-documents/verifiable-computing">Verifiable Computing</a></td><td></td></tr><tr><td><a href="/lilypad/research-and-vision/v1-research-documents/problem-statement">Problem Statement</a></td><td></td></tr><tr><td><a href="/lilypad/research-and-vision/v1-research-documents/prior-protocols">Prior Protocols</a></td><td></td></tr><tr><td><a href="/lilypad/research-and-vision/v1-research-documents/autonomous-agents">Autonomous Agents</a></td><td></td></tr><tr><td><a href="/lilypad/research-and-vision/v1-research-documents/mechanisms-to-explor">Mechanisms To Explore</a></td><td></td></tr></tbody></table>


# Verifiable Computing

### Verifiable Computing Basics

Verifiable computing is dedicated to ensuring that outsourced computations (that is, computations that are not done locally) are done correctly. In some scenarios, it cannot be assumed that the node to which a job is being outsourced will compute the result honestly, or that it is not faulty. Moreover, verifying the result should have less overhead than computing the result in the first place.

While blockchains provide safety and liveness, the massive replication of computation becomes too costly when that level of security is not needed. There is a difference between global consensus, which is necessary in blockchain environments, and local consensus, which is more suited for two-sided marketplaces. In global consensus, all nodes need to be convinced that every computation was done correctly. In contrast, in local consensus, only a small number of nodes - potentially only one node, the client - needs to be convinced that a computation was done correctly.

Ostensibly, for a two-sided marketplace, this implies that only a client really needs to be convinced that a computation was done correctly. However, these computations are not done in isolation, and the interrelation between a client choosing one node repeatedly versus many different nodes, and the mathematics behind those decisions, as well as the need to create a protocol that any client can come along to with no prior experience and trust that cheating is disincentivized, implies the creation of a global game that, while not requiring global consensus in the traditional sense, emulates it in some manner.

### Approaches to Verifiable Computing

#### Cryptographic Approaches

One way to ensure that computations were done correctly is by using cryptographic methods. There are a number of cryptographic approaches for verifiable computation, including

**Interactive Proof (IP)**

* In interactive proofs, verification of a statement is modeled as an interaction between a prover and a verifier. The goal of the prover is to convince the verifier that the statement is true, even when the verifier does not have the computation resources to do the computation itself.
* The protocol must satisfy completeness (if the statement is true, an honest verifier will be convinced) and soundness (if the statement is false, the prover cannot convince the verifier except with some negligible probability).

**Zero-Knowledge Proof (ZKP)**

* Zero-knowledge proofs are a type of interactive proof where the verifier learns nothing about private inputs of the computation, other than that the outputs were computed correctly from the all the inputs (some of which may be public/known to the verifier).
* A ZKP can be made non-interactive, in which case it is called a Non-Interactive Zero-Knowledge Proof (NIZK). Two common variants of NIZKs are zk-SNARKs (zero-knowledge Succinct Non-interactive Argument of Knowledge) and zk-STARKs (zero-knowledge Scalable Transparent Argument of Knowledge).
* Like IPs, ZKPs must also satisfy the requirements of completeness and soundness.

**Multi-Party Computation (MPC)**

* Multi-party computation allows multiple parties to jointly compute a function over their individual inputs without any party revealing its input to other parties. The main objectives of MPC are privacy (parties should learn known about each others' inputs), security (some level of anti-collusion preventing malicious attempts to learn information), functionality (the ability to compute functions over data), and robustness (the protocol should work correctly even in the presence of malicious behavior or faults).

#### Trusted Execution Environment (TEE)

* Trusted Execution Environments are secure and isolated enclaves, where code and data inside of the enclave are insulated from the rest of the system, including the operating system, applications, and other enclaves. The goal is to maintain both the confidentiality and the integrity of the code and data.

#### Verification-via-Replication

Verification-via-replication - often described using the adjective "optimistic" in the blockchain world - relies on recomputing the computation to check whether the end result is the same. The benefits of this method are that it is the easiest to understand, and in some sense, the easiest to implement.

In contrast to the other approaches, verification-via-replication often requires reliance on game-theoretic mechanisms such as collateral slashing, reputation, and other methods. This can become a bit complex when trying to counter collusion between the nodes that are computing the results.

One of the downsides of this approach is, of course, the extra effort expended on recomputing computations. However, with proper incentives, the overhead of this can be reduced dramatically. It is also important to keep in mind that the overhead of this approach is much lower than cryptographic methods, which usually have much higher overhead.

### Our Approach

We opt for verification-via-replication as a first approach, for the reasons that it is simple to understand, has less overhead than cryptographic approaches, and has an attack surface that can be economically modelled and analyzed.

This has the downside of making private computations difficult. While the inputs and outputs of jobs can be encrypted so that only the client and compute node can see the code and data, this still leaves the client vulnerable to having their information leaked. Future approaches can incorporate SLAs and eventually support for homomorphic encryption to deal with this issue.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fgit-blob-a73838ccf61fbb13efe7e58359aa6e58a10d7f05%2Fimage.png?alt=media" alt=""><figcaption><p>Current contract design</p></figcaption></figure>


# Problem Statement

The Distributed Compute Problem

### Problem Setup

The setup is a trustless, permissionless, two-sided marketplace for compute, where clients can purchase compute services from compute nodes. Trustless means that by default, the protocol does not assume that any particular node behaves in a trustworthy manner and that each node should be considered as rationally self-interested (note that this excludes intentionally malicious behavior). Permissionless means that any node can join or leave the network at will.

Matches for compute jobs are made off-chain, with the resulting deals and results recorded on-chain. Both clients and compute nodes need to agree to matches before they become deals, and make deposits to the protocol to enable rewards and punishments. Results are verified using verification-via-replication, and clients can check the results of any job after it has been completed, but before it needs to pay. It does so by calling upon a mediation protocol. The mediation protocol is the ultimate source of truth, and the outcome of the mediation protocol determines how payouts to nodes are made.

The issue of preventing fake results in the presence of a Trusted Third Party (TTP) as a mediator is effectively a solved problem (for example, see the section on prior verification-via-replication protocols, though there is much more literature on this topic). Given the assumption that the mediation protocol is the source of truth, we can treat the mediation protocol as a TTP. Since the fake results problem is basically already solved in this case, the cheating problem reduces down to solving the collusion problem within the mediation protocol. (Note, however, that we will address both cheating and collusion; the framework described here exists to conceptually simplify the problem.)

This is a typical scenario of a Byzantine environment, and we can use well-established approaches to Byzantine Fault Tolerance when invoking mediation. However, most BFT algorithms and cryptographic methods rely on assumptions regarding some fraction of honest nodes. The problem is that rational, utility-maximizing agents may still collude, even in a mediation consortium, in order to converge on incorrect results. On the one hand, we could assume that some fraction of nodes are honest, as is often done. On the other hand, can we do better?

## Problem Statement

### Task

The task is to find the mechanisms that incentivizes all nodes to behave honestly.

### Adversary model

All agents in the protocol are utility-maximizing. This will be elucidated in a subsequent section. Most of the literature has focused on the case where compute nodes are dishonest. However, the client can also behave dishonestly in some manner that maximizes their utility. For example, if the client has some level of control over the choice of mediator, and dishonest nodes have their collateral slashed, the client could collude with the mediator in order to deem a correct result incorrect and get a cut of the honest compute node's collateral.

### What is a good solution?

"Good" solutions can take a number of forms:

1. Nodes never have an incentive to be dishonest.
2. Nodes have an incentive to be dishonest that goes to zero as a function of the parameters of the protocol.
3. (1) or (2), but under some simplifying assumptions, such as there being some fraction of honest nodes within every mediation protocol.

A good solution would achieve any of these goals. Alternatively, another valuable outcome of this research would be to discover under what assumptions these goals can or cannot be met, or if the goals are even possible to achieve at all.

### Mechanisms for achieving these goals

There are a number of ways that these goals may be achieved. The approach will be to construct a digital twin of the protocol and test a number of different mechanisms in simulation. These mechanisms include variations on mediation protocols, collateralization, staking, taxes and jackpots, and others; see the Mechanisms to Explore section for details.


# Prior Protocols

Investigation of some previous verification-by-replication computing protocols

## Prior Verification-via-Replication Protocols

Before explaining our approach, we give a short overview of three prior approaches to verification-via-replication distributed computing protocols: Truebit, Modicum, and Smart Contract Counter-Collusion. We will outline potential improvements, and how our work builds on top of, and differs from, prior work.

### [Truebit](https://arxiv.org/pdf/1908.04756.pdf)

Truebit is a protocol for outsourcing computation from blockchains, built using smart contracts on top of Ethereum. The original potential use cases were trustless mining pools, trustless bridges, scaling transaction throughput, and, scalable “on-chain” storage. Since its goal is to scale on-chain computation, it aims for global consensus: "Since there exist no trusted parties on Ethereum’s network, by symmetry we must allow any party to be hired to solve any computational task, and similarly anyone should be able to challenge a Solver’s outcome of a computational task. The latter requirement ensures that TrueBit operates by **unanimous consensus**." (emphasis added)

The components of Truebit are Miners, Task Givers, Solvers, Verifiers, and Judges. In order to incentivize checking results, random errors are forced into computations, with jackpots awarded to those who find them. These jackpots are funded by taxes on computations.

The verification game consists of a series of rounds, where in each round, a smaller and smaller subset of the computation is checked. Eventually, only one instruction is used to determine whether a Solver or Verifier is correct: ["In fact, only one instruction line is used in a verification game. There will be a part of the program code where there is a discrepancy between the Solver and the Verifier. The instruction of that discrepancy point is used to verify who is right."](https://truebit.io/guide/truebit-structure/)

The authors claim that Sybil attacks are mitigated by pairwise Sybil-resistance between the parties of Task Givers, Solvers, and Verifiers, with Judges and Referees, whose roles are played by Miners, assumed to function as intended. Likewise, they claim that attacks to get bogus solutions on-chain by scaring off Verifiers are mitigated by the economics of deposits, taxes, and jackpot rewards. Additionally, a cartel of Solvers who absorb losses until they find a task with a forced error, upon which time they will receive the jackpot, will lose money in the long-term, since the expected Solver deposit per task is higher than the expected jackpot per task. Addressing another attack, the authors claim that an attack involving a flood of Challengers who try to take as much of the jackpot reward resulting from a forced error as possible is mitigated by having the total jackpot reward decrease as the number of Challengers increases.

#### Challenges

* Does not scale to large/complicated/arbitrary computations
* No formal theorems or proofs, no simulations, many plausible but unsubstantiated claims, especially regarding collusion
* Everything is done on-chain
* This model does not work well with two-sided marketplaces, because
  * It aims for global consensus, where any node is allowed to do the computation, whereas in two-sided marketplaces, clients need to be able to choose which nodes they are paying to do the computation
  * Clients may have time restrictions on their computations, and cannot wait for cases where their computations were injected with forced errors
* No accounting for repeated games

#### Takeaways

* Taxes and jackpots are a valuable tool to create a global game that affects local outcomes
* Provides a list of potential client attacks

### [Modicum](https://www.researchgate.net/profile/Aron-Laszka/publication/341640091_Mechanisms_for_Outsourcing_Computation_via_a_Decentralized_Market/links/5f14736b299bf1e548c3712a/Mechanisms-for-Outsourcing-Computation-via-a-Decentralized-Market.pdf)

The original version of [Modicum](https://www.researchgate.net/profile/Aron-Laszka/publication/341640091_Mechanisms_for_Outsourcing_Computation_via_a_Decentralized_Market/links/5f14736b299bf1e548c3712a/Mechanisms-for-Outsourcing-Computation-via-a-Decentralized-Market.pdf) had five key components: Job Creators (JC), Resource Providers (RP), Solvers (market makers), Mediators (agreed-upon third parties for mediation), and Directories (file systems, which we have replaced with IPFS and Docker registries). Job Creators are clients, the ones who have computations that need to be done and are willing to pay. Resource Providers are those with computational resources that they are willing to rent out for the right price. Solvers are market-makers; they match the offers from JCs and RPs. Mediators are third parties trusted by both JCs and RPs to arbitrate disagreements. The Directories are network storage services available to both JCs and RPs.

Job Creators are only allowed to submit deterministic jobs to the protocol. The Mediator exists to check for non-deterministic tasks submitted by the Job Creator (which can be used by the Job Creator as an attack vector to get free results), and fake results returned by the Resource Provider. The method for determining whether a job is deterministic or not is for the Mediator to run a job n times and check to see whether it receives different answers.

Modicum combines two separate ideas: checking the result from a Resource Provider to see if it is correct, and checking a result from a job submitted by a Job Creator to see if the job was deterministic or not. This implies that there is no capability for a client to simply check whether a result is correct or not, without the possibility of its collateral being slashed.

An alternative to trusting the Mediator (to act as a TTP) by having it run a job n times is having a consortium of n Mediators each run the task a single time. However, this adds the complication of achieving consensus in that consortium.

The issue of the Resource Provider and Mediator colluding to return a fake result is not addressed by this protocol. The authors allow for a Job Creator or Resource Provider to remove a Mediator from their list of trusted Mediators if they no longer trust it. However, that still leaves room to cheat at least once, and ideally this should be disincentivized from the outset.

There is also the issue of collateralization. The Modicum protocol, as well as a number of other protocols, assume (approximate) guesses as to the cost of jobs in advance, so that nodes can deposit the correct amount of collateral. However, doing so is fraught with complications; we provide an alternative approach in the Mechanisms to Explore section.

#### Challenges

* The Mediator is basically a trusted third party
* Client cannot simply check a result without being slashed, which is a consequence of the client attack model
* No accounting for repeated games

#### Takeaways

* Potential client attack, though one that can be mitigated by technical means
* The client has benefit of getting correct results, which needs to be accounted for in simulation
* Useful prototype for a two-sided marketplace (the [follow-up paper](https://par.nsf.gov/servlets/purl/10355144) for stream processing applications is also useful)

### [Smart Contract Counter-Collusion](https://arxiv.org/pdf/1708.01171.pdf)

The authors determine that cryptographic methods for verifiable computation are too expensive for real-world scenarios. For that reason, they rely on verification-via-replication. The scenario is one in which a client *simultaneously* outsources computation to two clouds, where those two clouds deposit collateral into smart contracts in such a way to create a game between them, where the game incentivizes doing the computation honestly. The central challenge that the authors tackle is the issue of collusion - that is, what if the two clouds collude on an incorrect answer?

In contrast to Modicum, the client is assumed to be honest, and in contrast to Truebit, a trusted third part (TTP) is used to handle disputes.

#### Three Contracts

The authors use a series of three contracts to counter collusion.

The first game is an induced Prisoner's Dilemma - to avoid the two clouds colluding, one cloud can be rewarded the other cloud's deposit (minus payment to the TTP for resolving the dispute) if the former returned the correct result and the latter did not. Thus, each cloud is better off giving the other cloud fake results while computing the correct result itself. This contract is called the **Prisoner's contract**. It is analogous to the equilibrium in the classic prisoner's dilemma being defection <> computing honest result and giving other node fake result if offered to collude.

However, the clouds can agree to collude via a smart contract as well. They could do this by both depositing another amount into a contract, where the leader of the collusion must add a bribe (less than its cost of computing) to the contract as well (disregarding the bribe, both clouds deposit the same amount of collateral). The deposit is such that the clouds have more of an incentive to follow the collusion strategy than to deviate from it. This contract is called the **Colluder's contract**.

In order to counteract the Colluder's contract, a **Traitor's contract** is used to avoid this scenario by incentivizing the clouds to report the Colluder's contract. The basic concept is that the traitor cloud indeed reports the agreed-upon collusion result to the client in order to avoid the punishment in the Colluder's contract, but also honestly computes and returns the result to the client in order to avoid the punishment of the Prisoner's contract. The client must also put down a deposit in the Traitor's contract. Only the first cloud to report the Colluder's contract gets rewarded. The signing and reporting of the contracts must happen in a particular order in order for this process to work.

The authors prove that these games individually and together lead to a sequential equilibrium (which is stronger than a Nash equilibrium), meaning that it is optimal not only in terms of the whole game, but at every information set (basically the set of options each player has at every turn).

#### Challenges

* A Colluder's contract can be signed on different chains (or even off-chain). In order to mitigate this, the Traitor's contracts would have to become cross-chain (which is a major technical challenge), not to mention the possibility of cryptographically secure contracts (e.g. MPC contracts) where one of the parties alone would not be able to prove the existence of this contract
* Relies on trusted third party to resolve disputes
* Every task is replicated (that is, two copies of each job are always computed)
* Assumes client is honest
* Assumes amount of collateral known beforehand
* No accounting for repeated games
  * It is well known that in the repeated Prisoner's dilemma, depending on the assumptions, cooperation becomes the equilibrium

#### Takeaways

* The contracts and the payoffs that they induce offer a valuable toolbox to think about the problem of collusion
* The contracts offer, in a restricted setting, an ironclad way (assuming the proofs are correct) of preventing collusion


# Autonomous Agents

Our Approach

### Utility Maximization

A core assumption in much of game theory is that agents are utility-maximizing. That is, agents are completely rational actors, and are able to execute exactly the behavior that maximizes their return, however "return" is defined in the given context.

However, we know that in real life, humans are not completely rational, and are not capable of perfect execution of actions. In that light, how can we look at the game-theoretic approaches in the last section?

Either we can try to account for the irrational behavior of humans, or we can try to emulate the behavior of utility-maximizing agents. While there is a large amount of game-theoretic literature dedicated to the former, we opt for the latter for reasons that will become clear below.

While this problem setting - verifiable computation by way of game theory - is different than many game theoretic settings, we can draw inspiration from commonly used concepts like the [revelation principle](https://en.wikipedia.org/wiki/Revelation_principle) and [strategyproofness](https://en.wikipedia.org/wiki/Strategyproofness). Both strategyproofness and the revelation principle are centered around the idea of incentivizing agents to truthfully report their preferences. Most approaches in the literature rely on analytic methods to determine what rational agents will do by analyzing their payoffs as a function of their preferences, the behaviors of other agents, and the mechanism under analysis. Ultimately, we are also aiming to find (a) mechanism(s) that lead(s) to an equilibrium where all agents choose to not cheat and not collude.

### Autonomous Agents

Note that the actual environment of a two-sided marketplace for distributed computation is extremely complicated (e.g. the heterogeneity of hardware, types of computation, latencies and bandwidths, etc.). Any theoretical/analytic approach to the problem that is actually correct should also work in simulation, so we opt for a simulation-driven approach.

The way that we can emulate perfectly rational behavior is by training autonomous agents to act on behalf of their human owners in a utility-maximizing manner. At that point, the challenge is to design the global game to drive the probability of cheating to zero - ideally, to make it be equal to zero - which is no small feat in a trustless and permissionless environment. However, the simplifying assumption that we are in fact operating with utility-maximizing agents conceptually simplifies the problem immensely.

The process begins by creating a digital twin of a two-sided marketplace. In this environment, autonomous agents acting on behalf of client and compute nodes will be trained to maximize returns based on data gathered in simulation. For now, we will avoid maximizing returns by optimizing scheduling, though this is a future topic of interest. We will use techniques primarily from the field of multi-agent reinforcement learning in order to train the agents. The precise methods we will use (e.g. modes of training and execution, homogeneous vs. heterogeneous agents, choice of equilibrium, self-play vs. mixed-play, value-based vs. policy-based learning, etc.) will be determined in the course of building the simulation. See the [pre-print](https://www.marl-book.com/) by Albrecht, Christianos, and Schäfer for our reference text.

At a minimum, the action space for an autonomous agent representing a compute node should be to cheat or not to cheat, and to collude or not collude within a mediation protocol. The observable environment for nodes on the network should include all data stored on the blockchain - that is, the sequence of deals, results, and mediations - as well as the information in the orderbook. While the orderbook will be off-chain, we model in the digital twin the orderbook acting as a centralized, single source of truth that all agents have access to. In the long-term, nodes will have (potentially non-identical) local information regarding other job and resource offers on the network.

Further work may explore agents forming beliefs about the hardware and strategies of other agents, but that is beyond the scope of the current work.

### First Principles Approach

We conclude with two "axioms" upon which we will base our simulations:

1. Every agent attempts to maximize its utility, including cheating and/or colluding if necessary.
2. All other components of the game should lead to a "good" solution, as defined in the problem statement.


# Mechanisms To Explore

## Mechanisms to Explore

The following is a list of mechanisms that we are currently considering exploring in order to mitigate attacks. Note that some of these mechanisms clearly would not be able to deter cheating and collusion alone. However, in combination with other mechanisms, they may achieve the goals. In this sense, they should be thought of as modules, optionally composable with each other.

### Mediation

Clients call upon a mediation protocol in order to verify the results of a node. There are several variations of the structure of the mediation protocol; the following parameters can be varied:

1. The number of nodes in the mediation protocol.
2. If more than two nodes in the mediation consortium, the consensus threshold that determines which result is the one considered to be correct.
3. How the nodes are chosen.
   * For example, we may want as a baseline the same constraint as in Modicum - that only mediators that both the client and the compute node mutually trust can be used for mediation.
     * Even with this baseline, there is still a question of how to choose the node(s) - it can be random, be determined by an auction, or any other method.
4. Recursive mediation - that is, if there is no consensus in the consortium, do another mediation.
   * There is a large space of possibilities regarding how to execute this.
   * There needs to be a limit to the number of nodes this recursive process can use. For example, the set of potential nodes can be the same as the set of mutually trusted mediators, as described above.
5. Other methods discussed here, such as taxes and jackpots, as well as staking and prediction markets, can be incorporated into the mediation protocol.

### Collateralization

There are a number of different types of collateral that need to be included in the protocol.

The client needs to deposit collateral so that the compute node knows that it can be paid. For computations where the cost is known up front, this is simple. However, it becomes complicated for arbitrary compute; the client might not have put up enough collateral initially, so there may have to be a back-and-forth between client and compute node where the latter halts the computation until the former deposits more collateral or simply cancels the computation and pays for the partially computed result. If the client does not pay, then the compute node can invoke a mediation process.

The compute node needs to deposit several types of collateral.

1. Collateral in case they timeout.
   * This is put up as part of the deal agreement - that is, when the deal is posted on-chain.
2. Collateral in case they cheat.
   * The way that the compute node will convey the amount of collateral they will deposit to indicate that they will not cheat is via a collateral multiplier. The compute node commits to a multiple of whatever they will charge the client ahead of time as part of the deal agreement. The actual collateral is put up after the result is computed and sent to the client. This simplifies immensely the task of determining how much collateral to deposit for arbitrary computations.
3. Collateral in case they don't do the computation at the rate they said they would. This is closely related to timeout collateral.
   * Ideally, this is a way of enforcing deadlines on jobs.
   * It is not necessary to make this collateral slashing binary - for example, a late result can be still be rewarded.
   * It is enforceable if, for example, the compute node says that they will do X WASM instructions/time. However, technical limitations may make this unrealistic, and it needs to be tested.

One possible way to overcome collusion is to require super high collateral for some particular nodes against each other that that even discount factors very favorable to collusion would not incentivize collusion, even when accounting for repeated games.

While this is not a part of anti-cheating mechanisms, collateral pooling could lower capital requirements for collateralization. High capital requirements are a second-order concern, but will become a higher priority once robust anti-cheating mechanisms are implemented.

### Taxes and Jackpots (inspiration from Truebit)

Taking inspiration from the taxes and jackpots scheme used in Truebit, deals can be taxed, with those taxes going to a jackpot that is then used to reward nodes via some distribution protocol determined by the mediation process. For this, we want to be able to take any fraction of the jackpot(s) and distribute it arbitrarily to arbitrary nodes (perhaps even those not involved in the mediation process).

This is a particularly interesting approach because the taxation + jackpots mechanism inherently create a global game that impacts local outcomes. While it may lead to potential collusion attacks, the tool alone is very useful, especially in conjunction with some other the other methods mentioned here. Modeling it in simulation would also provide the opportunity to test some of the hypotheses in the Truebit paper.

This method may also be useful in creating a robust platform where some clients do not care to check their results. That is, if some clients do not check results in general, it may be difficult to assert that the network is secure. Taxes and jackpots may be a way to address this.

### Prediction/Replication Markets

Prediction markets have been well-studied in a variety of different fields. More recently, a type of prediction market called a replication market has been explored in the context of replicability in science. With this inspiration, it may be possible that allowing nodes to make bets regarding the replicability of the computations of nodes may be useful in mitigating cheating. For example, nodes with a low prediction for replicability may act as a signal for that node's reputation and encourage it to behave honestly.

It is possible to overlay this mechanism on top of taxes, allowing nodes to choose where their taxes go in the prediction market.

Additionally, since Automated Market Makers are closely related to prediction markets, we can leverage many DeFi tools in this context.

### Staking behind nodes

Allow users to stake behind nodes. This is similar to prediction markets, but with slightly different economics. Like with prediction markets, it may be possible to tax users and then allow them to choose which nodes they stake behind. Overall, this approach is similar to delegated Proof-of-Stake.

### Announcing successful cheating

Can a node announcing that it successfully cheated (and thereby receiving a reward) benefit the robustness of the protocol? How much would this node have to be rewarded?

### Frequency of checks

How often should a client check results? Clearly it is related to the amount of collateral that the other node deposits, how much they value getting true/false results, reputation, and so on. This is a parameter that the client would need to learn to maximize its own utility.

### Reputation

The ledger can maintain a record, for each compute node, of the number of jobs the compute node has completed, the number of times its results were checked, and the number of times those results were replicated successfully. All other nodes (client and compute) could locally run some arbitrary function over these numbers to determine how reputable they find that node.

### Storing Inputs/Outputs

Results can only be replicated for as long as the inputs are stored somewhere. The client, compute node, or some other entity can pay for storing the inputs/outputs of jobs. The longer they are stored, the more time there is to check the results, which affects things like collateralization, the frequency of checks, etc.

This is related to, but not totally overlapping with, the amount of time that a node might have to wait before getting paid, which is the same time interval that a client has to check a result. However, checking the result after the node gets paid and receives back its collateral may still be possible, with other penalty schemes (or reward schemes, for example, coming from jackpots).

### Anti-Collusion via Obfuscation

Colluding requires the following knowledge in order to enforce the parameters of the collusion.

1. The public keys of the nodes participating in collusion.
2. The results that were posted by those public keys.
3. The payouts to the public keys.

In order to sign a collusion contract to begin with, the public keys must be known. However, in a mediation protocol with enough nodes, it may be possible to obscure (2) and (3) by

1. Having nodes submit results to the mediation protocol in an obscured/anonymous way
2. Have nodes be paid out according to the results of the mediation protocol in an obscured/anonymous way

If these two criteria can be met, then a mediation protocol based on them might be capable of imitating the game-theoretic outcomes seen in the Smart Contract Counter-Collusion paper.

There have been many decades of cryptography and security research focusing on similar problems to these. It may be the case that it is already possible to do this; otherwise, there is a large amount of ongoing research on the topics of privacy-preserving transactions, and much prior work in the flavor of secret-sharing/MPC/Tor/Monero/ZKPs that could enable this.


# Game Theory

Adding Trust to the Network

## Introduction

Lilypad is a verifiable trustless decentralized compute network that aims to prevent cheating in the network. The network consists of clients and compute nodes. The main goal is to establish a game theoretic approach to verifiable computing, where clients can trust the results they receive from compute nodes. The approach used in Lilypad is pure verification by replication, without relying on cryptographic tools like snarks or trusted execution environments.

## Global Consensus vs Local Consensus

In the context of Lilypad, it's important to understand the difference between global consensus and local consensus. Global consensus, as seen in blockchains, ensures that every node knows the result of a computation is correct due to massive replication across many nodes. However, in Lilypad's two-sided marketplace, only the client needs to be convinced that the result is correct. Local consensus is sufficient for the client, while other nodes like verifying nodes, requesters, and market makers may also need to be convinced.

## Adversary Model

Lilypad assumes non-malicious but utility-maximizing agents, including both clients and compute nodes. Utility-maximizing means that nodes will do anything to maximize their return, including returning false results if necessary. However, they are not malicious and won't purposely send back false results. The adversary model assumes that all nodes are utility-maximizing and does not consider malicious behavior.

## Good Solutions

Lilypad aims to achieve the best outcome, where nodes never have any incentive to cheat. If that's not possible, the goal is to minimize the incentive to cheat as a function of the protocol's parameters. Another option is to achieve the first two outcomes under simplifying assumptions, such as having a fraction of honest nodes in every mediation protocol. Lilypad also considers the possibility of not being able to achieve any of these outcomes and aims to understand the limitations.

## Reinforcement Learning Approach

Lilypad takes an adversary-first approach by designing the most powerful adversary possible and then optimizing against it. The team uses multi-agent reinforcement learning to train agents to act as utility-maximizing agents on behalf of clients and compute nodes. Reinforcement learning has shown impressive results in various domains, and Lilypad aims to leverage its capabilities to train agents that can maximize their utility in the network.

## Anti-Cheating Mechanisms

Lilypad plans to test various anti-cheating mechanisms once the reinforcement learning agents are implemented. These mechanisms include:

1. Consortium of mediators: A group of trusted mediators can check the results, reducing the likelihood of cheating.
2. Prediction markets and staking: Nodes can stake behind other nodes and lose their collateral if the node they stake behind is found to have cheated.
3. Taxes and jackpots: A tax is imposed on every job, and the taxes go into a jackpot that is awarded to nodes that find other nodes to have cheated.
4. Reputation: Nodes can build up a reputation based on the number of jobs they've done and the honesty of their results.
5. Sorting inputs and outputs: Storing inputs and outputs for longer periods allows for easier verification but increases collateralization requirements.
6. Frequency of checking results: Autonomous agents can decide when to check results, balancing the need for verification with the cost of collateralization.

## Future Developments

In the future, Lilypad aims to take into account the preferences of nodes, such as computational requirements, time requirements, and scheduling requirements. The long-term goal is to have compute nodes and clients negotiate with each other automatically over these aspects of a job. Lilypad acknowledges that game-theoretic verifiable computing is a less studied form of verifiable computing compared to zero-knowledge proofs and trusted execution environments. However, the team is committed to exploring this approach and conducting rigorous research to find effective solutions.

## Conclusion

Lilypad's game-theoretic approach to verifiable computing aims to prevent cheating in a decentralized compute network. By using reinforcement learning and testing various anti-cheating mechanisms, Lilypad strives to create a trustless environment where clients can have confidence in the results they receive from compute nodes. The team is actively working on implementing reinforcement learning agents and conducting simulations to evaluate the effectiveness of different strategies.


# Whitepaper

See [roadmap](https://lilypad.tech/#roadmap).&#x20;


# Meet the Team

Cross-functional, seasoned team

Lilypad is assembling a team of experts across a variety of domains to tackle and build product from  the complex R & D required for a decentralised compute network, including verification, privacy, token economics and more.

## Core Team

<table data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-type="files"></th><th data-hidden data-card-cover data-type="files"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>Alison Haire</td><td>CEO</td><td>prev: Filecoin, IBM, Accenture</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FAqpJYrNnPnbqlnM8rf5F%2Fimage.png?alt=media&amp;token=bd54facd-c63c-417f-9e24-61122430afb2">image.png</a></td><td></td></tr><tr><td>Steve King</td><td>COO</td><td>prev: Atlassian, Canva, Founder</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FdJFWPe0i4pYBnX6LgdQG%2Fimage.png?alt=media&amp;token=f8b64e29-3c9c-40a7-b888-89439170c2a2">image.png</a></td><td><a href="https://www.linkedin.com/in/austeveking/">https://www.linkedin.com/in/austeveking/</a></td></tr><tr><td>Stanley Bishop [PhD] </td><td>Head of Research</td><td>AI &#x26; ML Engineer &#x26; Researcher, Stanford University</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FB4xAgx4ElAfcnu85edEe%2Fimage.png?alt=media&amp;token=5efb9112-c6fb-4e21-8293-03aa9119bbf7">image.png</a></td><td><a href="https://www.linkedin.com/in/sciencestanley/">https://www.linkedin.com/in/sciencestanley/</a></td></tr><tr><td>James Walker</td><td>Engineering Manager</td><td>prev: Fission VP of Engineering</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fo3tlGVJ63nzur7WTd2Wj%2Fimage.png?alt=media&amp;token=2be621a2-715a-48e7-9fed-ec19112f24e8">image.png</a></td><td><a href="https://www.linkedin.com/in/walkah/">https://www.linkedin.com/in/walkah/</a></td></tr><tr><td>Eric Boniardi [PhD ML Research]</td><td>Senior Research Engineer [AI &#x26; ML, Systems Design]</td><td>Scholar - Mathematical Engineering</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FdoPLz5r1PFOLDqo7jhuO%2Fimage.png?alt=media&amp;token=8c6a3b83-1497-4c23-b9ef-b2c112fe3334">image.png</a></td><td><a href="https://www.linkedin.com/in/eric-boniardi/overlay/photo/">https://www.linkedin.com/in/eric-boniardi/overlay/photo/</a></td></tr><tr><td>Clint Snider [PhD Systems Engineering]</td><td>R &#x26; D Engineer [Tokenomics &#x26; Systems Design]</td><td>University of Florida, Tempest Labs</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FSdU3XIN9QLf7GnJ6oJxd%2Fimage.png?alt=media&amp;token=4b6bde77-1016-4e48-b258-bd82616208cb">image.png</a></td><td><a href="https://www.linkedin.com/in/clintonsnider/">https://www.linkedin.com/in/clintonsnider/</a></td></tr><tr><td>Lindsay Walker<br><br>Product Growth Lead</td><td>prev: Protocol Labs, Sauce Labs</td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FrHZM6xThPakHhH3JhWkb%2Flindsay.jpg?alt=media&amp;token=3f70b75b-07df-46cf-81e1-f21b943bf771">lindsay.jpg</a></td><td><a href="https://www.linkedin.com/in/lindsayjowalker/">https://www.linkedin.com/in/lindsayjowalker/</a></td></tr><tr><td>Nate Smith</td><td>Partnerships &#x26; BD Lead</td><td>Coinbase, prev: Gamestop, Godwoken</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FpqbxRB11SIYbsOwv366N%2Fimage.png?alt=media&amp;token=412f39d4-d9f7-4e33-9c19-a956ef7b2c00">image.png</a></td><td><a href="https://www.linkedin.com/in/natesmith12/">https://www.linkedin.com/in/natesmith12/</a></td></tr><tr><td>Brian Ginsburg </td><td>Software Engineer </td><td>prev: Fission</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FdWxwXFFSBauREVeH7W4E%2FbginsCrop.png?alt=media&amp;token=74c0fe5e-bc56-42d8-b5d5-5e1f8b3d6c6f">bginsCrop.png</a></td><td><a href="https://brianginsburg.com/">https://brianginsburg.com/</a></td></tr><tr><td>Ayush Kumar </td><td>Software Engineer</td><td>Apybara Founder, prev: Protocol Labs, MovableInk        </td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FibDTrw7l4J0EdcI1cADB%2Fimage.png?alt=media&amp;token=df659fad-d7b2-4079-8e37-29238b532083">image.png</a></td><td><a href="https://www.linkedin.com/in/kelindi/">https://www.linkedin.com/in/kelindi/</a></td></tr><tr><td>Arsen Yeremin</td><td>Solutions Architect [AI &#x26; Blockchain]</td><td>prev: TypherAI CTO</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F2KBCCq3gglcpIFfPdQn4%2Fimage.png?alt=media&amp;token=b4bd1cd5-ac21-46bc-a44a-0126b9a344cf">image.png</a></td><td><a href="https://www.linkedin.com/in/arsenyeremin/">https://www.linkedin.com/in/arsenyeremin/</a></td></tr><tr><td>Narbeh Shahnazarian</td><td>Solutions Architect</td><td>prev: Expedia</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fp2cIIEWy3goA5UNs4N7u%2Fimage.png?alt=media&amp;token=bd383a16-86ba-4f0b-8c56-5617c6c9135d">image.png</a></td><td><a href="https://www.linkedin.com/in/narbeh-shahnazarian-214a68a0/">https://www.linkedin.com/in/narbeh-shahnazarian-214a68a0/</a></td></tr><tr><td>Holly Grimm</td><td>Senior Protocol &#x26; Blockchain Engineer</td><td>prev: Gitcoin, Algovera</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F4NwccZObt4O3BceNTonW%2Fholly-g.jpg?alt=media&amp;token=19b29c24-59eb-41c0-bd49-213a23d59897">holly-g.jpg</a></td><td><a href="https://www.linkedin.com/in/hollygrimm/">https://www.linkedin.com/in/hollygrimm/</a></td></tr><tr><td>Logan Lentz</td><td>Software &#x26; Developer Relations Engineer</td><td>Leto, prev: AWS</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fq563tYnUdorZoN4l1d9I%2Fimage.png?alt=media&amp;token=0fde5a79-2591-42d2-aeb1-c4e43d7f9a2b">image.png</a></td><td><a href="https://www.linkedin.com/in/logan-lentz/">https://www.linkedin.com/in/logan-lentz/</a></td></tr><tr><td>Alex Mirran</td><td>BD &#x26; DevRel Lead</td><td>Leto, prev: Willis Towers Watson, Disrupt FinTech</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F1iQ1ujvT3nMSVLXvry3C%2Fimage.png?alt=media&amp;token=7c39d7ce-5fee-4770-af00-e85e1bcd58de">image.png</a></td><td><a href="https://www.linkedin.com/in/alex-mirran/">https://www.linkedin.com/in/alex-mirran/</a></td></tr><tr><td>Phil Billingsby</td><td>Senior Developer Relations &#x26; DX Engineer </td><td>Unboxed, prev: Arweave, FilmFreeway</td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FouvBhl0j5ZRc8X1M7hNm%2Fimage.png?alt=media&amp;token=cc319b41-1be2-420e-a623-c0c714228afb">image.png</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FF5b4NocjjGMDQcL96S5G%2FPhil.jpg?alt=media&amp;token=02cfeb1c-1b3c-4132-9aec-8f2b02a7f45d">Phil.jpg</a></td><td><a href="https://www.linkedin.com/in/phil-billingsby-68a103303/">https://www.linkedin.com/in/phil-billingsby-68a103303/</a></td></tr></tbody></table>

## Advisors

<table data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th><th data-hidden data-card-cover data-type="files"></th></tr></thead><tbody><tr><td>David Aronchick</td><td>Distributed Compute &#x26; Product Advisor</td><td>Experience: Kubernetes, Bacalhau, Expanso CEO</td><td><a href="https://www.linkedin.com/in/aronchick/">https://www.linkedin.com/in/aronchick/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FxfJ7ox8yQPG9ftQb4unG%2Fimage.png?alt=media&amp;token=19561f94-da9f-4f7d-bcd6-d11bbdefa5f8">image.png</a></td></tr><tr><td>Peter Wang</td><td>AI &#x26; ML Advisor</td><td>Experience: Anaconda.ai Founder</td><td><a href="https://www.linkedin.com/in/pzwang/">https://www.linkedin.com/in/pzwang/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FRGWWZ21bXkGXoBwmNnPX%2Fimage.png?alt=media&amp;token=28c45c74-d13e-4578-b175-2a335589b628">image.png</a></td></tr><tr><td>Jonathon Victor</td><td>Web3 Landscape and Business Advisor</td><td>Experience: Ansa Research, Protocol Labs</td><td><a href="https://www.linkedin.com/in/jonathan-victor-69a44255/">https://www.linkedin.com/in/jonathan-victor-69a44255/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FVtA3Bb34spL7cXBWPUHL%2Fimage.png?alt=media&amp;token=4c448537-5e46-466c-820a-c43341dbeaf8">image.png</a></td></tr><tr><td>Paul Wagner</td><td>Business &#x26; Technical Advisor</td><td>Experience: Filecoin, Oracle, Startup Founder</td><td><a href="https://www.linkedin.com/in/thpaulwagner/">https://www.linkedin.com/in/thpaulwagner/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FDzU9GNXVgbBpntjEc5Cr%2Fimage.png?alt=media&amp;token=da933fc3-b7c2-407e-8817-98c46f551aa6">image.png</a></td></tr><tr><td>Brooklyn Zelenka</td><td>Technical &#x26; Business Advisor</td><td>Experience: Co-Founder Fission</td><td><a href="https://www.linkedin.com/in/brooklynzelenka/">https://www.linkedin.com/in/brooklynzelenka/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FjQ9V4X0RcSk2DAvebzEo%2Fimage.png?alt=media&amp;token=143f6efc-0048-449d-be4e-c8f80891bbda">image.png</a></td></tr><tr><td>Eric Vander Wal</td><td>Product &#x26; Web3 Advisor</td><td>Experience: Uprising Founder, Godwoken Founder, Oasys</td><td><a href="https://www.linkedin.com/in/eric-vander-wal/">https://www.linkedin.com/in/eric-vander-wal/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FpVJTJDM5yDwOLigu9XM0%2Fimage.png?alt=media&amp;token=47f1752a-3dd0-4be9-bf70-7db4c1ec0684">image.png</a></td></tr><tr><td>Jay Logelin</td><td>Engineering and Cryptography Advisor</td><td>Zama, prev: Protocol Labs, NEAR</td><td><a href="https://www.linkedin.com/in/jlogelin/">https://www.linkedin.com/in/jlogelin/</a></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FVdyPIP9FHHEXwXj5yKms%2F1707499922587.jpg?alt=media&amp;token=fe9a60e3-dfa4-4c2d-9723-e053c91bebbb">1707499922587.jpg</a></td></tr></tbody></table>

## Other Partners

<table data-view="cards"><thead><tr><th></th><th></th><th></th><th data-hidden data-card-cover data-type="files"></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td>Crypto Econ Lab</td><td>Tokenomics Design Team</td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fxb6mxqnBXalhu58i7p5Z%2Fcryptoeconlab_logo.jpeg?alt=media&amp;token=f7c0b03f-16bc-4d48-b99e-f5d3f26417ab">cryptoeconlab_logo.jpeg</a></td><td><a href="https://cryptoeconlab.io/">https://cryptoeconlab.io/</a></td></tr><tr><td>Frontline.codes<br><br>Frontend Development Team<br><br></td><td></td><td></td><td><a href="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FHsbGbSTiX2IsRCeF5MPL%2FScreenshot%202024-10-25%20124214.png?alt=media&amp;token=7cd01d41-39cf-4ed9-a3ec-cb77d13b1d77">Screenshot 2024-10-25 124214.png</a></td><td><a href="https://frontline.codes/">https://frontline.codes/</a></td></tr></tbody></table>


# Events

Catch up with the Lilypad team or hack on the Lilypad Network at these events!

## Upcoming

View the upcoming events we are attending or hosting.

### **2025**

| Event                                                           | Location | Date                 |
| --------------------------------------------------------------- | -------- | -------------------- |
| [Filecoin AI Blueprints](https://www.encode.club/ai-blueprints) | Online   | March 17th           |
| [TOKEN2049 Dubai](https://www.dubai.token2049.com/)             | Dubai    | April 30th - May 1st |

## Past

View the previous events we attended or hosted.

### 2025

| Event                                   | Location | Date                      |
| --------------------------------------- | -------- | ------------------------- |
| [ETHDenver](https://www.ethdenver.com/) | Denver   | February 23rd - March 2nd |

### **2024**

<table><thead><tr><th width="249">Event</th><th width="204">Location</th><th>Date</th></tr></thead><tbody><tr><td><a href="https://devcon.org/en/">Devcon</a></td><td>South East Asia</td><td>November 9th - November 17th</td></tr><tr><td><a href="https://ethcc.io/">ETHCC Brussels</a></td><td>Brussels</td><td>July 8th - July 11th</td></tr><tr><td><a href="https://ethglobal.com/events/hackfs2024">HackFS</a></td><td>Online</td><td>May 17th - June 7th</td></tr><tr><td><a href="https://ethglobal.com/events/sanfrancisco2024">ETH Global</a></td><td>San Francisco</td><td>October 18th - October 20th</td></tr><tr><td><a href="https://devcon.org/">Devcon</a></td><td>Bangkok</td><td>November 12th - November 15th</td></tr><tr><td><a href="https://ethglobal.com/events/sydney">ETHGlobal</a></td><td>Sydney</td><td>May 3rd - May 5th</td></tr><tr><td><a href="https://twitter.com/protocollabs/status/1694728785189540063?s=20">LabWeek</a></td><td>Istanbul</td><td>November 13th - November 17th</td></tr><tr><td><a href="https://fildev.io">Fil:Dev Iceland</a></td><td>Iceland</td><td>September 25h - September 28th</td></tr><tr><td><a href="https://smartcon.chain.link/">Chainlink SmartCon</a></td><td>Barcelona</td><td>October 1st - October 3rd</td></tr><tr><td><a href="https://fil-vegas.io">Fil:Vegas</a></td><td>Las Vegas</td><td>October 3rd - October 5th</td></tr><tr><td><a href="https://twitter.com/Lilypad_Tech/status/1688319601107521536?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm%5E1688319601107521536%7Ctwgr%5E2f6c362765cea6018c3db4cc5bce5e710baa50fd%7Ctwcon%5Es1_&#x26;ref_url=https%3A%2F%2Fcdn.iframe.ly%2FodriGT3%3Fapp%3D1">Open Data Hack</a></td><td>Online</td><td>September 23rd</td></tr></tbody></table>


# Run a node - Linux

Instructions for setting up a Resource Provider (node) on the public Lilypad testnet using Linux, including obtaining necessary funds, installing required software, and ensuring security measures.

{% hint style="warning" %}
These instructions are no longer actively maintained. We recommend using the [Docker approach](/lilypad/resource-providers/run-a-node) for the most up-to-date and supported setup.
{% endhint %}

## Prerequisites

{% hint style="info" %}
For a more in-depth look at the requirements to run a Lilypad node, please refer to the [hardware requirements](/lilypad/resource-providers/hardware-requirements) documentation.
{% endhint %}

* Linux (*Ubuntu 22.04* LTS)
* Nvidia GPU
* [Nvidia drivers](https://ubuntu.com/server/docs/nvidia-drivers-installation)
* [Docker](https://docs.docker.com/engine/install/ubuntu/) (Ubuntu install)
* Nvidia Docker drivers

{% hint style="info" %}
Lilypad RPs currently only support Linux installs. Running a RP on Windows is currently experimental. The resource provider version of Lilypad is not supported on Darwin/macOS.
{% endhint %}

## Network information and testnet tokens

The testnet has a base currency of ETH, as well as a utility token called LP. Both are used for running nodes. To add a node to the testnet, follow these steps:

### Metamask

We recommend using MetaMask with custom settings to make things easier. Once you have it installed and setup, here are the settings you need to use:

{% hint style="info" %}
Network name: Arbitrum Sepolia

New RPC URL: <https://sepolia-rollup.arbitrum.io/rpc>

Chain ID: 421614

Currency symbol: ETH

Block explorer URL: (leave blank)
{% endhint %}

For a step by step guide on adding the network and importing the LP testnet token, please refer to our [Setting up MetaMask documentation](/lilypad/quickstart/cli/setting-up-your-wallet).

### Fund your wallet with ETH and LP

To obtain testnet LP, use the [Lilypad faucet](http://faucet.lilypad.tech) and enter your ETH address.

To obtain testnet ETH, use a third party [Arbitrum Sepolia testnet faucet](https://arbitrum.faucet.dev/ArbSepolia) and enter your ETH address.

{% hint style="info" %}
The Arbitrum Sepolia faucet provides 0.0001 tokens per request. If you need more tokens and already have Sepolia ETH, you can use the [official Arbitrum bridge](https://bridge.arbitrum.io/) to transfer it over to Arbitrum Sepolia.
{% endhint %}

The faucet will give you both ETH (to pay for gas) and LP (to stake and pay for jobs).

## Installation

To set up your environment for using Lilypad with GPU support, you need to install several key components. This guide will walk you through installing Docker, the Nvidia Container Toolkit, Bacalhau, and Lilypad. You'll also configure systemd to manage these services efficiently.

### Install Docker

Docker is a platform that allows you to automate the deployment of applications inside lightweight, portable containers.

To install Docker Engine, follow the steps specific to your operating system from the official Docker documentation:

* [Linux](https://docs.docker.com/engine/install/ubuntu/) - Docker Engine

### Install Nvidia Container Toolkit

To ensure proper operation of your graphics cards and Lilypad, follow these steps to install the Nvidia Toolkit Base Installer: [Nvidia Container Toolkit download page](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html)

```bash
curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
  && curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \
    sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
    sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
    
sudo apt-get update

sudo apt-get install -y nvidia-container-toolkit
```

Configure the container runtime by using the `nvidia-ctk` command:

```bash
 sudo nvidia-ctk runtime configure --runtime=docker --set-as-default
```

The `nvidia-ctk` command modifies the `/etc/docker/daemon.json` file on the host. The file is updated so that Docker can use the NVIDIA Container Runtime.

Restart the Docker daemon:

```bash
sudo systemctl restart docker
```

### Install Bacalhau

Bacalhau is a peer-to-peer network of nodes that enables decentralized communication between computers. The network consists of two types of nodes, which can communicate with each other.

{% hint style="info" %}
Lilypad now supports Bacalhau `v1.6.0`&#x20;
{% endhint %}

To install Bacalhau, run the following in a new terminal window (run each command one by one):

```bash
cd /tmp

wget https://github.com/bacalhau-project/bacalhau/releases/download/v1.6.0/bacalhau_v1.6.0_linux_amd64.tar.gz

tar xfv bacalhau_v1.6.0_linux_amd64.tar.gz

sudo mv bacalhau /usr/bin/bacalhau

sudo chown -R $USER /app/data
```

To check your Bacalhau version use:

```bash
bacalhau version
```

The expected output is:

```
CLIENT  SERVER  LATEST
v1.6.0  v1.6.0  <latest Bacalhau version>
```

If the Bacalhau `CLIENT` version is not v1.6.0, it will need to be replaced. Follow the steps [here](https://docs.lilypad.tech/lilypad/hardware-providers/troubleshooting#how-do-i-change-my-bacalhau-version) to uninstall and reinstall Bacalhau.

### Install Lilypad

The installation process for the Lilypad CLI involves several automated steps to configure it for your specific system. Initially, the setup script identifies your computer's architecture and operating system to ensure compatibility. It will then download the latest production build of the Lilypad CLI directly from the official GitHub repository using `curl` and `wget`.

Once the CLI tool is downloaded, the script sets the necessary permissions to make the executable file runnable. It then moves the executable to a standard location in your system's path to allow it to be run from any terminal window.

**Via official released binaries**

```bash
# Detect your machine's architecture and set it as $OSARCH
OSARCH=$(uname -m | awk '{if ($0 ~ /arm64|aarch64/) print "arm64"; else if ($0 ~ /x86_64|amd64/) print "amd64"; else print "unsupported_arch"}') && export OSARCH
# Detect your operating system and set it as $OSNAME
OSNAME=$(uname -s | awk '{if ($1 == "Darwin") print "darwin"; else if ($1 == "Linux") print "linux"; else print "unsupported_os"}') && export OSNAME;
# Remove existing lilypad installation if it exists
sudo rm -f /usr/local/bin/lilypad
# Download the latest production build
curl https://api.github.com/repos/lilypad-tech/lilypad/releases/latest | grep "browser_download_url.*lilypad-$OSNAME-$OSARCH-gpu" | cut -d : -f 2,3 | tr -d \" | wget -i - -O lilypad
# Make Lilypad executable and install it
chmod +x lilypad
sudo mv lilypad /usr/local/bin/lilypad
```

{% hint style="info" %}
Ensure the RP is running the latest Lilypad version. Two weeks after each version release, the previous Lilypad version is deprecated.
{% endhint %}

To verify the installation, run `lilypad` in the terminal to display the version and a list of available commands, indicating that Lilypad CLI is ready to use.

### Write env file

You will need to create an environment directory for your node and add an environment file that contains your node's private key.

To do this, run the following in your terminal:

```bash
sudo mkdir -p /app/lilypad
sudo touch /app/lilypad/resource-provider-gpu.env
```

Next, add your node's private key into `/app/lilypad/resource-provider-gpu.env`:

```bash
WEB3_PRIVATE_KEY=<YOUR_PRIVATE_KEY> (the private key from a NEW MetaMask wallet FOR THE COMPUTE NODE)
```

{% hint style="info" %}
A [`WEB3_PRIVATE_KEY`](https://docs.lilypad.tech/lilypad/lilypad-milky-way-testnet/install-run-requirements#id-2.-set-web3_private_key) can be retrieved from the Metamask account details menu. For more info, check out [this](https://support.metamask.io/managing-my-wallet/secret-recovery-phrase-and-private-keys/how-to-export-an-accounts-private-key/) official guide from Metamask on viewing a wallet's private key.
{% endhint %}

This is the key where you will get paid in LP tokens for jobs run on the network.

{% hint style="warning" %}
You must not reuse your compute node key as a client, even for testing: this will result in failed jobs and will negatively impact your compute node since the wallet address is how nodes are identified on the network.
{% endhint %}

### Setup Arbitrum RPC

The Lilypad Network uses the Arbitrum Sepolia Testnet to settle compute transactions. When a transaction is ready to be saved on-chain, Lilypad cycles through a list of public Arbitrum Sepolia RPC endpoints using the endpoint that settles first to save the compute transaction.

Resource Providers have the option to [setup their own Arbitrum RPC endpoint](https://docs.lilypad.tech/lilypad/hardware-providers/setup-arbitrum-rpc) using Alchemy instead of using the default public RPC endpoints.

A personal RPC endpoint helps RPs to avoid reliability issues with the public RPC endpoints used by Lilypad ensuring rewards can be earned and jobs can be run consistently. RPs running a personal RPC endpoint contribute to the fault tolerance and decentralization of the Lilypad Network! Read more in the Alchemy Arbitrum [docs](https://docs.alchemy.com/reference/arbitrum-api-quickstart).

### Install systemd unit for Bacalhau

systemd is a system and service manager for Linux operating systems. systemd operates as a central point of control for various aspects of system management, offering features like parallelization of service startup, dependency-based service management, process supervision, and more.

To install systemd, open `/etc/systemd/system/bacalhau.service` in your preferred editor:

```bash
sudo vim /etc/systemd/system/bacalhau.service
```

```
[Unit]
Description=Lilypad V2 Bacalhau
After=network-online.target
Wants=network-online.target systemd-networkd-wait-online.service

[Service]
Environment="LOG_TYPE=json"
Environment="LOG_LEVEL=debug"
Environment="HOME=/app/lilypad"
Environment="BACALHAU_SERVE_IPFS_PATH=/app/data/ipfs"
Restart=always
RestartSec=5s
ExecStart=/usr/bin/bacalhau serve --orchestrator --compute

[Install]
WantedBy=multi-user.target  
```

### Install systemd unit for GPU provider

Open `/etc/systemd/system/lilypad-resource-provider.service` in your preferred editor.

{% hint style="info" %}
Hint: `sudo vim /etc/systemd/system/lilypad-resource-provider.service`
{% endhint %}

```
[Unit]
Description=Lilypad V2 Resource Provider GPU
After=network-online.target
Wants=network-online.target systemd-networkd-wait-online.service

[Service]
Environment="LOG_TYPE=json"
Environment="LOG_LEVEL=debug"
Environment="HOME=/app/lilypad"
Environment="OFFER_GPU=1"
EnvironmentFile=/app/lilypad/resource-provider-gpu.env
Restart=always
RestartSec=5s
ExecStart=/usr/local/bin/lilypad resource-provider 

[Install]
WantedBy=multi-user.target
```

Reload systemd's units/daemons (you will need to do this again if you ever change the systemd unit files that we wrote, above)

```bash
sudo systemctl daemon-reload
```

## Start Lilypad node

Start systemd units:

{% hint style="info" %}
Start bacalhau and lilypad-resource-provider in a different terminal window than the ipfs node.
{% endhint %}

{% hint style="info" %}
Once bacalhau is started, wait 5-20 seconds before starting lilypad-resource-provider. Bacalhau sometimes takes time to start and lilypad-resource-provider will not start without bacalhau.
{% endhint %}

```bash
sudo systemctl enable bacalhau
sudo systemctl enable lilypad-resource-provider
sudo systemctl start bacalhau
sudo systemctl start lilypad-resource-provider
```

Now that your services have been installed and enabled, check the status of Bacalhau to ensure it is running correctly on your node:

```bash
sudo systemctl status bacalhau
```

## View node status

To check if the node is running use the following command:

```bash
sudo systemctl status lilypad-resource-provider
```

This will give a live output from the Lilypad node. The logs will show the node running and accepting jobs on the network.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FzbYFmQ1uMx7rQtLw37Ti%2Fcarbon.png?alt=media&amp;token=ce3c8436-d042-4c5d-a6af-f79cbaf4fb22" alt=""><figcaption></figcaption></figure>

Run the following command to get more status info from your node:

```bash
sudo journalctl -u lilypad-resource-provider.service -f
```

To restart your resource provider run:

```bash
sudo systemctl restart lilypad-resource-provider
```

## Support for Lilypad RPs

For **complex issues, bug reports, or feature requests,** open a discussion in the Lilypad-Tech Github organization discussion [board](https://github.com/orgs/Lilypad-Tech/discussions).

* Navigate to the discussion [board](https://github.com/orgs/Lilypad-Tech/discussions), select "New Discussion", choose "rp-issues", and fill out the template.
* Without a discussion opened, our team will not be able to support the problem.

**For quick questions or minor issues,** use the Lilypad Discord [#i-need-help](https://discord.com/channels/1212897693450641498/1230231823674642513) channel and provide the following info.

* Description (including Lilypad version running on your node)
* Hardware Info (including Linux/Windows version)
* Related blockchain/ETH addresses of transaction hashes
* Output Logs - `sudo systemctl status lilypad-resource-provider`
* Related links/urls
* Screenshots

## Update Lilypad version

When a new version of Lilypad is released, it is important for resource providers to update their installations to ensure compatibility and access to the latest features and improvements.

Please note that using `sudo rm -rf` is very powerful and can be dangerous if not used carefully.

1. If the Lilypad RP is running, stop the system (if the node is not running, disregard this first step):

```bash
sudo systemctl stop bacalhau
sudo systemctl stop lilypad-resource-provider
```

2. Remove the Lilypad executable by running:

```bash
sudo rm -rf /usr/local/bin/lilypad
```

3. [Reinstall Lilypad with the latest version](https://docs.lilypad.tech/lilypad/hardware-providers/run-a-node/linux#install-lilypad)
4. Start your resource provider by running:

```bash
sudo systemctl start bacalhau
sudo systemctl start lilypad-resource-provider
```

## Disconnecting a node

To disconnect your node from Lilypad you will need to do a few things to completely offboard.

{% hint style="warning" %}
Using `sudo rm -rf` is very powerful and can be dangerous if not used carefully. It is highly recommended to navigate to the parent directory and remove files from there to avoid accidental deletion of important files.
{% endhint %}

First, stop the node:

```bash
sudo systemctl stop bacalhau
sudo systemctl stop lilypad-resource-provider
```

Next, you must remove the `.service` files related to Lilypad and Bacalhau. These files are typically stored in `/etc/systemd/system/`. To remove them, run the following command:

```bash
sudo rm -rf /etc/systemd/system/lilypad-resource-provider.service /etc/systemd/system/bacalhau.service
```

Next we notify the systemd manager to reload its configuration by running:

```bash
sudo systemctl daemon-reload
```

Then, remove the environment file for the Lilypad resource provider. This file is usually stored in `/app/lilypad/`. To remove it, run:

```bash
sudo rm -rf /app/lilypad/resource-provider-gpu.env
```

Finally, if you followed the installation instructions from the Lilypad documentation and moved the executable to `/usr/local/bin/lilypad`, it can be removed from there. If the executable is stored in a different directory on your machine, navigate to that directory and remove it from there. To remove the executable, run:

```bash
sudo rm -rf /usr/local/bin/lilypad
```

To remove Bacalhau, run:

```bash
sudo rm -rf /usr/bin/bacalhau
```

As every system is different, these instructions may vary. If you have any issues, please reach out to the team in the [Lilypad Discord](https://lilypad.team/discord) for help!

## View Lilybit\_ rewards

To view your Lilybit\_ rewards, visit one of the following dashboards and paste your node's public address into the input:

* [Grafana dashboard](https://grafana.lilypad.tech/d/adxhou3o1q8sga/rewards-per-wallets?orgId=1\&refresh=1m)
* [Lilypad leaderboard](https://info.lilypad.tech/leaderboard)

## Security

If you want to allowlist only certain modules (e.g. Stable Diffusion modules), to control exactly what code runs on specific nodes (which can be audited to ensure that they are secure and will have no negative impact on the nodes), set an environment variable `OFFER_MODULES` in the GPU provider to a comma separated list of module names, e.g. `sdxl:v0.9-lilypad1,stable-diffusion:v0.0.1.`

Visit the [Lilypad GitHub](https://github.com/Lilypad-Tech/lilypad#available-modules) for a full list of available modules.

## Run a node video guide

{% embed url="<https://www.youtube.com/watch?v=slCLg35rAU4>" %}


# Setup Arbitrum RPC

Personal RPC endpoint for a Lilypad Resource Provider

The Lilypad Network uses the Arbitrum Sepolia Testnet to settle compute transactions.

{% hint style="info" %}
The Lilypad Network is not responsible for issues caused by misconfigured RPC endpoints. Please use at your own risk!
{% endhint %}

This personal RPC endpoint allows Resource Providers (RP) to avoid reliability issues with the RPC endpoints used by Lilypad ensuring rewards can be earned and jobs can be run consistently. RPs running a personal RPC endpoint contribute to the fault tolerance and decentralization of the Lilypad Network! Read more in the Alchemy Arbitrum [docs](https://docs.alchemy.com/reference/arbitrum-api-quickstart).

## Setup RPC video guide

{% embed url="<https://www.youtube.com/watch?v=INKpdrfF0Xs>" %}


# Alchemy RPC

Setup a personal Arbitrum Sepolia RPC for a Lilypad Resource Provider.

This guide walks through the steps of setting up a personal RPC endpoint for Arbitrum Sepolia using [Alchemy](https://www.alchemy.com/).

{% hint style="info" %}
This guide demonstrates how to use Alchemy's RPC, but this can be substituted for any Arbitrum Sepolia RPC that supports WebSockets.
{% endhint %}

## Setup Alchemy account

[Create an account](https://dashboard.alchemy.com/) and login to the Alchemy dashboard.

Select the “free” tier as the compute units provided should be sufficient to run a Lilypad RP. The free service provides 300 million compute units per month.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FjT7xDtKeNZo1hvxMYMqo%2FScreenshot%202024-09-17%20at%2010.39.32%E2%80%AFAM.png?alt=media&amp;token=3a742f43-aede-42ca-83a7-f36b8054c95f" alt=""><figcaption></figcaption></figure>

Select “skip bonus” or input a credit card with billing info (the card will not be charged unless the compute credits in the free tier are used).

## Setup RPC endpoint for Arbitrum Sepolia

In the “Overview” section of the Alchemy dashboard, navigate to “My app” and select “Endpoints”. If an app was not created upon login, create a new one by selecting "Create new app".

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FA2vpnEyLonYsVXPfv2wL%2Falchemy1-1.png?alt=media&amp;token=a8033cde-2bd2-4e92-a6a6-aa3719046e9d" alt=""><figcaption></figcaption></figure>

By selecting “Endpoints”, the “Networks” tab will open providing an option to configure the Arbitrum API.

* Select “Sepolia”
* Select “Websockets”

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FVOUuB6JLHv2phAEJaegp%2Falchemy1.png?alt=media&amp;token=78a707c6-653f-4c38-8662-7f5db4e76cc4" alt=""><figcaption></figcaption></figure>

The RPC endpoint for Arbitrum Sepolia is ready to be used with the Lilypad Resource Provider:

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FZbS4JiQMFZYBcN033W4j%2Falchemy3.png?alt=media&amp;token=4ed7b18d-a71c-4a62-bd22-d080347a19cc" alt=""><figcaption></figcaption></figure>

Metrics for the RPC can be viewed in the “Metrics” tab.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FtvPkzEX33O81dM1GGvuc%2FScreenshot%202024-09-17%20at%2011.00.45%E2%80%AFAM.png?alt=media&amp;token=15a85bc5-551e-4514-ba16-9fd3a29fccce" alt=""><figcaption></figcaption></figure>

## Use the new RPC endpoint

{% hint style="info" %}
This is guide is for individuals running a Lilypad Resource provider, find more info [here](https://docs.lilypad.tech/lilypad/hardware-providers/run-a-node).
{% endhint %}

Lilypad RPs can use a personal RPC endpoint with a few simple steps. Only Web-socket (WSS) connections are supported.

### **Docker users**

Stop the existing Lilypad Resource Provider (RP) before setting up the new RPC.&#x20;

Locate the Lilypad RP Docker container using:

```
docker ps
```

Stop the container using the PID:

```
docker stop <container ID>
```

Use this command to start the lilypad-resource-provider.service with the new RPC:

```
docker run -d --gpus all -e WEB3_PRIVATE_KEY=<private-key> -e WEB3_RPC_URL=wss://arb-sepolia.g.alchemy.com/v2/some-id-from-alchemy --restart always ghcr.io/lilypad-tech/resource-provider:latest
```

Check the status of the container:

```
docker logs <container ID>
```

### **Ubuntu users**

Stop the existing Lilypad RP (if the node is not running, disregard this first step):

```
sudo systemctl stop bacalhau
sudo systemctl stop lilypad-resource-provider
```

Update `lilypad-resource-provider.service` with the new RPC:

```
sudo nano /etc/systemd/system/lilypad-resource-provider.service
```

Add following line to \[Service] section:

```
Environment="WEB3_RPC_URL=wss://arb-sepolia.g.alchemy.com/v2/some-alchemy-id"
```

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FxM9mNOcfbcMvOYywseS1%2Fimage%20(26).png?alt=media&amp;token=14d16707-18d2-46bf-b2f3-1b9e5cb8a896" alt=""><figcaption></figcaption></figure>

Reboot the node:

```
sudo reboot
```

If the Lilypad RP was [setup](https://docs.lilypad.tech/lilypad/hardware-providers/run-a-node/linux#install-systemd-unit-for-bacalhau) properly as a systemd service, the RP will reboot using the new RPC. Once the reboot is complete, the RP should be running with the updated configuration. To verify your node is back online and running correctly, run the following:

```
sudo systemctl status lilypad-resource-provider
sudo systemctl status bacalhau
```


# Infura RPC

Setup a personal Arbitrum Sepolia RPC for a Lilypad Resource Provider.

This guide walks through the steps of setting up a personal RPC endpoint for Arbitrum Sepolia using [Infura](https://www.infura.io/).

{% hint style="info" %}
This guide demonstrates how to use Infura's RPC, but this can be substituted for any Arbitrum Sepolia RPC that supports WebSockets.
{% endhint %}

## Setup Infura account

[Create an account](https://app.infura.io/login) on Infura and choose your plan based on how many APIs you need.

Select the “free” tier as the compute units provided should be sufficient to run a Lilypad RP.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2FsIMOVxSXEM4qHvzX2YWg%2FScreenshot%202024-10-14%20at%2011.20.56%E2%80%AFAM.png?alt=media&amp;token=bce7595a-b268-4947-a443-fc6b3d3705bf" alt=""><figcaption></figcaption></figure>

## Setup RPC endpoint for Arbitrum Sepolia

In the Infura dashboard, a new API key will usually generate automatically. If not, select "Create New API Key". Navigate to "Configure" to setup the API key.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fr0PUGhXxo1La7atc5qK2%2Finfuraguide1.png?alt=media&amp;token=c7e4f475-e6c7-437f-8311-ba913525856a" alt=""><figcaption></figcaption></figure>

Scroll down the list to the Arbitrum network and ensure the Sepolia testnet box is checked, then save changes.

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F6DILKX7Cb2Sdmkns8bTb%2FScreenshot%202024-10-14%20at%2011.28.38%E2%80%AFAM.png?alt=media&amp;token=410865ed-1963-43dd-9842-ee7d3e01bda6" alt=""><figcaption></figcaption></figure>

In the API key dashboard, select "Active Endpoints" and navigate to "WebSockets".

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2Fz7o0e0NdvS3ropQgdjaH%2Finfuraguide2.png?alt=media&amp;token=107e6d67-bc2d-407d-bef8-30973f180d87" alt=""><figcaption></figcaption></figure>

Scroll down the page to find the Arbitrum Sepolia URL. The RPC endpoint for Arbitrum Sepolia is ready to be used with the Lilypad Resource Provider:

<figure><img src="https://58337262-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FtadiyoOe4nTUoSulEVOV%2Fuploads%2F9jPpqWX7W6e9IKqW1f9N%2Finfuraguide3.png?alt=media&amp;token=c1c20f42-433e-4940-870c-63c2ff2f8289" alt=""><figcaption></figcaption></figure>

## Use the new RPC endpoint

{% hint style="info" %}
This is guide is for individuals running a Lilypad Resource provider, find more info [here](https://docs.lilypad.tech/lilypad/hardware-providers/run-a-node).
{% endhint %}

Lilypad RPs can use a personal RPC endpoint with a few simple steps. Only Web-socket (WSS) connections are supported.

### **Docker users**

Stop the existing Lilypad Resource Provider (RP) before setting up the new RPC.&#x20;

Locate the Lilypad RP Docker container using:

```
docker ps
```

Stop the container using the PID:

```
docker stop <container ID>
```

Use this command to start the lilypad-resource-provider.service with the new RPC:

```
docker run -d --gpus all -e WEB3_PRIVATE_KEY=<private-key> -e WEB3_RPC_URL=wss://arbitrum-sepolia.infura.io/ws/v3/some-id-from-infura --restart always ghcr.io/lilypad-tech/resource-provider:latest
```

Check the status of the container:

```
docker logs <container ID>
```

### **Ubuntu users**

Stop the existing Lilypad RP (if the node is not running, disregard this first step):

```
sudo systemctl stop bacalhau
sudo systemctl stop lilypad-resource-provider
```

Update `lilypad-resource-provider.service` with the new RPC:

```
sudo nano /etc/systemd/system/lilypad-resource-provider.service
```

Add following line to \[Service] section:

```
Environment="WEB3_RPC_URL=wss://arbitrum-sepolia.infura.io/ws/v3/some-id-from-infura"
```

Reboot the node:

```
sudo reboot
```

If the Lilypad RP was [setup](https://docs.lilypad.tech/lilypad/hardware-providers/run-a-node/linux#install-systemd-unit-for-bacalhau) properly as a systemd service, the RP will reboot using the new RPC. Once the reboot is complete, the RP should be running with the updated configuration. To verify your node is back online and running correctly, run the following:

```
sudo systemctl status lilypad-resource-provider
sudo systemctl status bacalhau
```


# Running Lilypad in a Front End

Setting up a feeless frontend for Lilypad

### What We'll Do <a href="#heading-what-well-do" id="heading-what-well-do"></a>

In this tutorial, we will review how to create a feeless front end for a Lilypad module. To do so, we will:

1. Create a new NextJS project.
2. Create a CLI Wrapper to run the module.
3. Create a server to handle requests from the browser.
4. Create our user interface to allow a user to enter a prompt to run the module.
5. Run the front end.

## Install Next JS <a href="#heading-install-next-js" id="heading-install-next-js"></a>

Create a new Next JS project by running the following:

```bash
npx create-next-app@latest
```

## Prepare your CLI Wrapper <a href="#heading-prepare-your-cli-wrapper" id="heading-prepare-your-cli-wrapper"></a>

In your project folder, create a new file named .env to manage your environment variables. Paste in the following, and replace “\<your-private-key-here>” with the private key of the wallet that will be charged for all of the jobs.

```plaintext
WEB3_PRIVATE_KEY=<your-private-key-here>
```

{% hint style="danger" %}
To avoid publishing your private key online, add .env to your `.gitignore` file.
{% endhint %}

Install `dotenv` to access your environment variables:

```bash
npm install dotenv
```

In your app folder, create a folder for pages, then a folder for API, and in that folder, create a file named `cliWrapper.js`. Paste the following code into `cliWrapper.js`:

```javascript
// Dotenv will allow us to access our environment variables so we can access our private key.
require('dotenv').config({ path: __dirname + '/../../../.env' });
// This will allow us to run code in our CLI.
const { exec } = require('child_process');

// The function we will call on the front end, to run a lilypad job.
function runCliCommand(userInput, callback) {
    console.log("Lilypad Starting...");
    // Ensure the WEB3_PRIVATE_KEY environment variable is set
    const web3PrivateKey = process.env.WEB3_PRIVATE_KEY;
    // If the private key was not set up properly, we should expect to see this error.
    if (!web3PrivateKey) {
        console.error('WEB3_PRIVATE_KEY is not set in the environment variables.');
        return;
    }

    // This command will first export our private key, and then run the Lilypad SDXL module with the prompt provided by the user.
    const command = `export WEB3_PRIVATE_KEY=${web3PrivateKey} && lilypad run sdxl-pipeline:v0.9-base-lilypad3 -i PromptEnv=Prompt="${userInput}”;

    // This is a callback function to handle any errors when calling runCliCommand function.
    exec(command, async (error, stdout, stderr) => {
        if (error) {
            console.error(`Error: ${error.message}`);
            return callback(error);
        }
        if (stderr) {
            console.error(`Stderr: ${stderr}`);
            return callback(stderr);
        }

        // When Lilypad runs successfully, it returns the relative path to the files it generated, and the IPFS url. Here we are grabbing the relative path in local storage to serve the image to our front end.
        const lines = stdout.trim().split('\n');
        const path = lines[lines.length - 4].trim(); // Trim any extra whitespace
        const filePath = path.replace('open ', '') + '/outputs/image-42.png';

        // This console log will confirm that Lilypad ran successfully.
        console.log(stdout)

        // This will return our output to the front end.
        if (callback) {
            callback(null, filePath);
        }
    });
}

module.exports = { runCliCommand };
```

## Prepare your server <a href="#heading-prepare-your-server" id="heading-prepare-your-server"></a>

Install Express and CORS:

```plaintext
npm install express cors
```

In your app folder, create a file named `server.js` and paste the following code:

```javascript
// Express is a framework for APIs.
const express = require('express');
// Cors will allow our browser to interact with our command line.
const cors = require('cors');
// Here we are importing our cli wrapper we just created
const { runCliCommand } = require('./pages/api/cliWrapper');
// Create an instance of Express to set up the server
const app = express();
// Define the port number on which the server will listen
const port = 3001;

// Enable CORS middleware to handle cross-origin requests
app.use(cors());
// Use express.json middleware to parse JSON requests
app.use(express.json());
// Use express.urlencoded middleware to handle URL-encoded data (useful for form submissions)
app.use(express.urlencoded({ extended: true }));

// Here we will serve images from the directory Lilypad saves files to/
app.use('/images', express.static('/tmp/lilypad/data/downloaded-files'));

// When a user submits their prompt, this will run our CLI Wrapper with their prompt.
app.post('/api/cliWrapper', (req, res) => {
    // Grab the user's input
    const userInput = req.body.userInput;
    // Run our CLI Wrapper with their input
    runCliCommand(userInput, (error, filePath) => {
        if (error) {
            return res.status(500).send('Error processing command');
        }
        // Convert file path to URL
        const urlPath = filePath.replace('/tmp/lilypad/data/downloaded-files', '');
        // Send the URL where the result can be accessed
        res.send(`http://localhost:${port}/images${urlPath}`);
    });
});

// Start the server and listen on the specified port
app.listen(port, () => {
    console.log(`Server listening at http://localhost:${port}`);
});
```

## Prepare your front end <a href="#heading-prepare-your-front-end" id="heading-prepare-your-front-end"></a>

In the App folder, delete the default content on the `Page.tsx` file and paste in the following code:

```typescript
"use client";
import React, { useState, FormEvent } from 'react';
import Image from 'next/image';

export default function Home(): JSX.Element {
  // State for storing the user's input
  const [userInput, setUserInput] = useState<string>('');

  // State for storing the URL of the generated image
  const [output, setOutput] = useState<string>('');

  // State to track loading status
  const [isLoading, setIsLoading] = useState<boolean>(false);

  // Function to handle the form submission
  const handleSubmit = async (event: FormEvent<HTMLFormElement>): Promise<void> => {
      event.preventDefault();
      setIsLoading(true); // Start loading
      try {
          const response = await fetch('http://localhost:3001/api/cliWrapper', {
              method: 'POST',
              headers: {
                  'Content-Type': 'application/json',
              },
              body: JSON.stringify({ userInput }),
          });
          const imageURL = await response.text();
          setOutput(imageURL);
      } catch (error) {
          console.error('Error:', error);
      }
      setIsLoading(false); // End loading
  };

  return (
    <main>
        <div className="flex min-h-screen flex-col items-center p-24">
            <div className="z-10 max-w-5xl w-full flex flex-col items-center justify-center font-mono text-sm space-y-4">
                <h1>Lilypad Image Generator</h1>
                <div className="w-[500px] flex flex-col items-center"> {/* Container for image and form */}
                    {/* Placeholder or Loading Animation */}
                    {!output && (
                        <div className="border-dashed border-2 border-gray-300 w-full h-[500px] flex items-center justify-center">
                            {isLoading ? (
                                <span className="text-gray-500">Loading...</span> // Replace with your loading animation
                            ) : (
                                <span className="text-gray-500">Image will appear here</span>
                            )}
                        </div>
                    )}
                    {/* Display the generated image if the output URL is available */}
                    {output && <Image src={output} alt="Generated" width={500} height={500} />}
                    <form onSubmit={handleSubmit} className="w-full mt-2">
                        {/* Textarea for user input */}
                        <textarea
                            className="w-full p-2"
                            value={userInput}
                            onChange={(e) => setUserInput(e.target.value)}
                            placeholder="Enter your prompt"
                            rows={4}
                        />
                        <div className="w-full flex justify-end">
                            {/* Submit button for the form */}
                            <button type="submit" className="mt-4">Submit</button>
                        </div>
                    </form>
                </div>
            </div>
        </div>
    </main>

  );
}
```

## Running your front end <a href="#heading-running-your-front-end" id="heading-running-your-front-end"></a>

To run your front end on your local host, you will need to update the `next.config.js` file. Open the file and paste the following:

```javascript
/** @type {import('next').NextConfig} */
const nextConfig = {
    images: {
      domains: ['localhost'],
    },
  }

module.exports = nextConfig;
```

Now you're ready to test your front end!

Open two terminals, one to run your server and another to run your front end. In both terminals, navigate to your App folder. For your server terminal, run the following:

```bash
node server.js
```

You should see “Server listening at [http://localhost:3001”](http://localhost:3001/).

In your second terminal, run your front end with the following command:

```bash
npm run dev
```

You should see something like this.

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1702326339073/1842bc6b-6741-4857-83e3-f1ce4f97d718.png?auto=compress,format\&format=webp)

In your browser, navigate to <http://localhost:3000/> and you should now be able to see your front end. It should look like this.

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1702487184879/f491f5ed-9649-4b50-b2fd-8fa08668db3e.png?auto=compress,format\&format=webp)

Type in a prompt and click on the enter button. If you look at the terminal running your server, you’ll notice that it says “Lilypad Starting…”. Lilypad will print to the console once the job is complete, and it should look something like this.

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1702326448561/e4845124-6683-4a63-a723-f82e81066432.png?auto=compress,format\&format=webp)

Now check your browser, and you should see your image.

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1702487323238/319c18e7-5a2f-4c9e-8854-3b798c6ab9f5.png?auto=compress,format\&format=webp)

Congrats! You have successfully integrated Lilypad into your front end.

### Share your custom front ends with us! <a href="#heading-share-your-custom-front-ends-with-us" id="heading-share-your-custom-front-ends-with-us"></a>

If you were able to complete the tutorial, share your front end with us on [Twitter](https://twitter.com/Lilypad_Tech)!

If you need any help or have any feedback, hop into [our Discord](https://lilypad.team/discord)!

View the [source code](https://github.com/Lilypad-Tech/lilypad-front-end).


# Lilypad Smart Contracts

Sorry! We're working on this page!

## Using the Lilypad Contracts

We have deployed the [LilypadOnChainJobCreator](https://github.com/Lilypad-Tech/lilypad/blob/main/hardhat/contracts/LilypadOnChainJobCreator.sol) contract which you can use to trigger running jobs on the lilypad network from other smart contracts.

It works in tandem with the `lilypad jobcreator` on-chain which will watch the on-chain contract and manage jobs on behalf of contracts that submit them.

### **Creating a job**

You will need to know the contract address for the on-chain job creator so we can submit transactions to it.

The production controller address is `0x8e136587e3e5266d5244f6aa896E5CAf8E969946` and you can ask it for the address of the on-chain job creator `getJobCreatorAddress()`

Running a job involves 2 phases:

* calling `approve` on the ERC-20 contract to allow the solver to spend your tokens
* trigger the job via the on chain job manager

Now we know the address of the on-chain job controller - we can ask it for 3 things:

* the address of the ERC-20 token contract - `getTokenAddress()`
* how much the required deposit it - `getRequiredDeposit()`
* the address of the solver that will handle running the job for us - `getControllerAddress()`

Knowing these 3 things means we can call the standard ERC-20 `approve` to allow the solver to spend our tokens on our behalf.

Now - we can call the `runJob` method of the on chain controller from another contract. This will cause the job-creator service to kick in and do the following things:

* check that funds have been approved for the solver
* transfer those funds to it's wallet
* run the job on lilypad
* call the `submitResults` method on the on-chain job creator
* the on-chain job creator will call the `submitResults` of the original calling contract

The following is an example on-chain smart contract:

```solidity
// SPDX-License-Identifier: Apache-2.0
pragma solidity ^0.8.6;

import "@openzeppelin/contracts/access/Ownable.sol";
import "@openzeppelin/contracts-upgradeable/proxy/utils/Initializable.sol";
import "./ILilypadJobManager.sol";
import "./ILilypadJobClient.sol";

contract ExampleClient is Ownable, Initializable, ILilypadJobClient {

  address private jobManagerAddress;
  ILilypadJobManager private jobManagerContract;

  mapping(uint256 => string) private jobResults;

  event JobCreated(
    uint256 id,
    string message
  );

  event JobCompleted(
    uint256 id,
    string dealId,
    string dataId
  );

  function initialize(address _jobManagerAddress) public initializer {
    setJobManagerAddress(_jobManagerAddress);
  }

  function setJobManagerAddress(address _jobManagerAddress) public onlyOwner {
    require(_jobManagerAddress != address(0), "Job manager address");
    jobManagerAddress = _jobManagerAddress;
    jobManagerContract = ILilypadJobManager(jobManagerAddress);
  }

  function getJobResult(uint256 _jobID) public view returns (string memory) {
    return jobResults[_jobID];
  }

  function runCowsay(
    string memory message
  ) public {
    string[] memory inputs = new string[](1);
    inputs[0] = string(abi.encodePacked("Message=", message));
    uint256 id = jobManagerContract.runJob(
      "cowsay:v0.0.4",
      inputs,
      msg.sender
    );

    emit JobCreated(
      id,
      message
    );
  }

  function submitResults(
    uint256 id,
    string memory dealId,
    string memory dataId
  ) public override {
    jobResults[id] = dataId;
    emit JobCompleted(
      id,
      dealId,
      dataId
    );
  }
}
```

Here is an example of a script that brings all of this together:

```javascript
import bluebird from 'bluebird'
import {
  connectToken,
  connectJobManager,
  connectExampleClient,
  getWallet,
  getAddress,
} from '../utils/web3'
import { ethers } from 'hardhat'

async function main() {
  // it's annoying to not be able to use argv but hardhat complains about it
  const message = process.env.MESSAGE || 'Hello World!'

  const token = await connectToken()
  const manager = await connectJobManager()
  const client = await connectExampleClient()

  const setRequiredDepositTx = await manager
    .connect(getWallet('solver'))
    .setRequiredDeposit(ethers.parseEther("2"))
  await setRequiredDepositTx.wait()

  const requiredDeposit = await manager.getRequiredDeposit()

  console.log(`requiredDeposit: ${Number(requiredDeposit)}`)

  const paytokensTx = await token
    .connect(getWallet('job_creator'))
    .approve(getAddress('solver'), requiredDeposit)
  await paytokensTx.wait()

  console.log(`tokens approved: ${paytokensTx.hash}`)

  const runjobTx = await client
    .connect(getWallet('job_creator'))
    .runCowsay(message)
  const receipt = await runjobTx.wait()
  if(!receipt) throw new Error(`no receipt`)

  console.log(`submitted job: ${runjobTx.hash}`)

  let jobID = 0

  receipt.logs.forEach(log => {
    const logs = client.interface.parseLog(log as any)
    if(!logs) return
    jobID = Number(logs.args[0])
  })

  console.log(`Job ID: ${jobID}`)
  console.log(`Waiting for job to be completed...`)

  let result = ''

  while(!result) {
    result = await client.getJobResult(jobID)
    if(!result) {
      await bluebird.delay(1000)
    }
  }

  console.log(`Job result: ${result}`)
}

main().catch((error) => {
  console.error(error);
  process.exitCode = 1;
});

```

### View results

After the job is completed, a job result ID will be displayed in your terminal. The job result can be accessed through the following URL:

```
https://ipfs.io/ipfs/<JOB RESULT ID>/stdout
```

{% hint style="info" %}
Please note that the IPFS link might take some time to display the result.
{% endhint %}

## Lilypad Hardhat template

The [Lilypad Hardhat template](https://github.com/Lilypad-Tech/lilypad-hardhat-template) demonstrates how to use Lilypad within the Hardhat development environment for onchain interactions. It provides the tools and structure needed to deploy smart contracts that interact with Lilypad nodes onchain.

This template includes an example of running a job using the [`cowsay`](/lilypad/developer-resources/module-marketplace/build-a-job-module/cowsay) module. The `cowsay` module is a simple program that generates an ASCII picture of a cow with a message.

### Getting started

You can run this template directly in [Gitpod](https://gitpod.io/new/#https://github.com/Lilypad-Tech/lilypad-hardhat-template).

The environment will automatically run a preinstall script that checks if Lilypad and Bacalhau are installed. If they are not already present on the user's machine, the script will install them.

1. Open the [Gitpod](https://gitpod.io/new/#https://github.com/Lilypad-Tech/lilypad-hardhat-template) environment
2. Run Lilypad:

   ```bash
   npm run lilypad:up
   ```

This script builds Docker services, starts necessary components (Bacalhau, geth), funds the accounts used, deploys the contracts, and sequentially launches Lilypad related services in the background.

### Usage

This will execute [cowsay](https://docs.lilypad.tech/lilypad/lilypad-milky-way-examples/hello-cow-world) onchain using Lilypad, with the specified input message.

In a separate terminal, run this command to execute cowsay onchain:

```bash
npm run lilypad:onchain:cowsay
```

To view the results of the job, refer to [this section](#view-results).

To update the smart contract, you can modify `ExampleClient` in `contracts/ExampleClient.sol`. Once you've made the changes, run `npm run deploy`.


# Older Versions


# Lilypad v0 (deprecated)


# Reference


# Lilypad v0 Quick Start

Get Started with Lilypad v0 - Call off-chain distributed compute from your smart contract!


# Deployed Network Details

Lilypad v0 Integrated Networks


# Lilypad Smart Contracts

The Lilypad v0 Smart Contracts


# Creating your own Jobs

Creating Bacalhau Job Spec's


# Architecture

Lilypad v0 is a "bridge" for running Bacalhau compute jobs via a smart contracts.


# Examples


# Stable Diffusion

Running a Stable Diffusion Job from a smart contract with Lilypad v0


# Drand Off-chain Voting

Implementing an unbiased off-chain voting process using Lilypad and Bacalhau, with Drand Time-lock encryption


# Ethereum Address Fraud Detector

Running a fraud detection job with Lilypad


# Lilypad v1 (deprecated)


# Reference


# Overview

Trustless Distributed Compute for web3


# Architecture Diagrams


# Get started here!!


# Setting up Metamask


# Funding your Wallet from Faucet


# \[CLI] Install Run Requirements


# \[CLI] Run "Hello, World!" Job


# \[Smart Contract] Run "Hello, World!" Job


# Run "Hello, World" Node




---

[Next Page](https://docs.lilypad.tech/lilypad/llms-full.txt/1)