AIGENTLAYER is an open-source project that aims to simplify interaction with EigenLayer's restaking capabilities. Instead of learning complex commands or navigating multiple interfaces, you can just tell AIGENT what you want in plain English.
- GaiaNet Integration: AIGENT uses GaiaNet as its default AI LLM API.
- Check TVL: Quickly check the total value locked in EigenLayer.
- Check operator status: Quickly check the status of an operator.
- Check staker status: Quickly check the status of a staker.
- Deposit stETH into EigenLayer: Request a deposit, and AIGENT will handle the transaction.
- Web Client: AIGENTLAYER comes with a web client that allows you to interact with the agent.
- Multiple Client Support: AIGENTLAYER supports multiple clients, including Discord, Twitter, Telegram, and more.
- Workflow Automation: AIGENT can automate workflows, such as restaking, claiming rewards, and withdrawing.
- EigenExplorer Full Integration: Gain real-time insights and data from the EigenLayer ecosystem.
- Advanced Evaluators: Intelligent decision-making—AIGENT can suggest when to restake, claim, or withdraw based on market conditions.
- Framework: ai16z Eliza
- Languages: Primarily TypeScript (Node.js, pnpm).
- Architecture: Modular monorepo allowing for easy extension of actions, providers, and evaluators.
AIGENTLAYER uses GaiaNet as its default AI LLM API. This integration simplifies how the agent interprets natural language, offering scalable and reliable language processing to power its autonomous actions. By building on GaiaNet’s infrastructure, we also align with the Gaia Agent Infrastructure Challenge criteria, demonstrating how integrating GaiaNet’s capabilities can enhance agent performance and contribute to the Gaia ecosystem’s growth.
Key files and folders:
.
├── assets/ # Assets
├── docs/ # Documentation
├── eliza/ # The AI Agent Folder
│ ├── agent/ # The AI Agent
│ ├── characters/ # Characters Folder
│ │ └── aigent.character.json # Character file
│ ├── packages/ # Packages Folder
│ │ ├── plugin-eigenlayer/ # Custom plugin for EigenLayer
│ │ │ └── src/ # Source code for the plugin
│ │ │ └── actions/ # Actions for the plugin
│ │ │ │ └── eigendeposit.ts # Action to deposit stETH into EigenLayer
│ │ │ └── evaluators/ # Evaluators for the plugin
│ │ │ └── providers/ # Providers for the plugin
│ │ │ │ └── tvl.ts # TVL data provider
│ │ │ │ └── operator.ts # Operator data provider
│ │ │ │ └── staker.ts # Staker data provider
│ │ └── client/ # Web Client
└── README.md # This file
Before getting started with Eliza, ensure you have:
- Python 2.7+
- Node.js 23+
- pnpm 9+
- Git for version control
- A code editor (VS Code or VSCodium recommended)
- CUDA Toolkit (optional, for GPU acceleration)
git clone https://github.com/worksgoodcompany/aigentlayer-demo.git- Move to agent folder
cd eliza- Install Dependencies
pnpm install- Copy the .env.example file and rename it to .env
cp .env.example .env4.1 Configure Environment (REQUIRED)
Edit .env and add your values:
EVM_PRIVATE_KEY= # Your Ethereum private key (needs to have testnet ETH and stETH)
INFURA_API_KEY= # Your Infura API key4.2 Configure Environment (OPTIONAL)
Edit .env and add your values:
# Suggested quickstart environment variables
# For Discord integration
DISCORD_APPLICATION_ID= # For Discord integration
DISCORD_API_TOKEN= # Bot token
# For Twitter integration
TWITTER_USERNAME= # Account username
TWITTER_PASSWORD= # Account password
TWITTER_EMAIL= # Account email
# For Telegram integration
TELEGRAM_BOT_TOKEN=
# For OpenAI API
OPENAI_API_KEY= # OpenAI API key
# For Grok API
GROK_API_KEY= # Grok API key
# For Anthropic API
ANTHROPIC_API_KEY= # Anthropic API key
# For OpenRouter API
OPENROUTER_API_KEY=
OPENROUTER_MODEL= # Default: uses hermes 70b/405b
# Gaianet Configuration
GAIANET_MODEL=llama
GAIANET_SERVER_URL=https://llama8b.gaia.domains/v1
SMALL_GAIANET_MODEL= # Default: llama3b
SMALL_GAIANET_SERVER_URL= # Default: https://llama3b.gaia.domains/v1
MEDIUM_GAIANET_MODEL= # Default: llama
MEDIUM_GAIANET_SERVER_URL= # Default: https://llama8b.gaia.domains/v1
LARGE_GAIANET_MODEL= # Default: qwen72b
LARGE_GAIANET_SERVER_URL= # Default: https://qwen72b.gaia.domains/v1
GAIANET_EMBEDDING_MODEL=
USE_GAIANET_EMBEDDING= # Set to TRUE for GAIANET/768, leave blank for local
# For EVM
EVM_PRIVATE_KEY=
INFURA_API_KEY=- Configure the character file (OPTIONAL)
Set up the clients you want to use and the AI model provider
"name": "AIGENT", # Change this to your desired name.
"plugins": [],
"clients": [], # Add the clients you want to use like ["discord", "twitter", "telegram"]
"modelProvider": "gaianet", # You can replace this with "openai", "openrouter" or others.
- Build the Agent
pnpm build- Run the Agent
pnpm run start --characters="characters/aigent.character.json"- Move to client folder
cd client
# or
cd eliza/client- Run the client
pnpm run dev- Access the client You can access the client using a web browser at: http://localhost:5173/
- Via web: Open the provided local URL in your browser.
- Via Telegram, Discord, or Twitter: Follow the instructions in
docs/clients.md(to be created).
What is eigenlayer?
Can you explain what restaking is?
What's the current TVL in EigenLayer?
Show operator status for 0x0a3e3d83c99b27ca7540720b54105c79cd58dbdd
Show staker status for 0x0000000f65d503603782d94e78e30c6d05955741
Please deposit 0.01 stETH into EigenLayer.
We welcome contributions! Submit pull requests for bug fixes, feature enhancements, or new evaluators.
MIT License