focEliza 🤖

Motivation

Enter ‘decentralized autonomous AI agent‘ !!!

This is an AI agent completely free from human administrator control, governed entirely by decentralized logic!

We’re not here to define what changes it will bring to the world—let it happen first! Let the AI agent live on-chain!

What do we build upon Eliza?

📖 Existing Work:

• AI16Z's Eliza: The first AI agent operating system.
• ElizaTEE: Implemented by Phala Network, achieving verifiable and confidential Eliza operations.

📅 Our Next Steps:

• Add Verifiable Runtime Proof Modules: Enable Eliza to sign logs, proving its actions to external parties for third-party verification.
• Add On-Chain Components: Bring Eliza AI agents to life on-chain, integrating their role systems, memory and keystores modules with on-chain states and DA (Data Availability).
• Upgrade TEE Node to TEE Cluster: Alongside on-chain components, allow AI agents to restore data, memory, and keystores from the chain and continue running on any TEE machine, enabling decentralized execution.
• Add Blockchain Interaction Components: Include trusted blockchain clients, data synchronizers, and other components running within the TEE to enable seamless interaction with blockchain systems.
• Integration into evm compatible networks
• Integration into Artela evm++ Layer1 networks and EVM++ Rollkit

🔧 Ongoing Features:

• plugin-tee-verifiable-log: While Eliza operates within the TEE, it uses a derived key pair to sign its actions, ensuring that these actions are definitively executed by Eliza. Third-party users can remotely verify Eliza's public key to validate these actions.
• plugin-tee-onchain-da: Eliza writes the "life" information of a specified AI agent—such as character files, memory, and keystore—into the blockchain (or DA layer) in near real-time. If the TEE node running the specified AI agent shuts down for any reason, another TEE node can download the AI agent's "life" data from the chain and resume its operation. This data is encrypted by the TEE, ensuring that only the same version of TEE Eliza can download and restore the agent.

Eliza 🤖

📑 Technical Report | 📖 Documentation | 🎯 Examples

🌍 README Translations

中文说明 | 日本語の説明 | 한국어 설명 | Persian | Français | Português | Türkçe | Русский | Español | Italiano | ไทย | Deutsch | Tiếng Việt | עִברִית | Tagalog | Polski | Arabic | Hungarian | Srpski | Română | Nederlands | Ελληνικά

🚩 Overview

✨ Features

• 🛠️ Full-featured Discord, X (Twitter) and Telegram connectors
• 🔗 Support for every model (Llama, Grok, OpenAI, Anthropic, Gemini, etc.)
• 👥 Multi-agent and room support
• 📚 Easily ingest and interact with your documents
• 💾 Retrievable memory and document store
• 🚀 Highly extensible - create your own actions and clients
• 📦 Just works!

Video Tutorials

AI Agent Dev School

🎯 Use Cases

• 🤖 Chatbots
• 🕵️ Autonomous Agents
• 📈 Business Process Handling
• 🎮 Video Game NPCs
• 🧠 Trading

🚀 Quick Start

Prerequisites

• Python 2.7+
• Node.js 23+
• pnpm

Note for Windows Users: WSL 2 is required.

Use the Starter (Recommended)

git clone https://github.com/elizaos/eliza-starter.git
cd eliza-starter
cp .env.example .env
pnpm i && pnpm build && pnpm start

Manually Start Eliza (Only recommended if you know what you are doing)

Checkout the latest release

# Clone the repository
git clone https://github.com/elizaos/eliza.git

# This project iterates fast, so we recommend checking out the latest release
git checkout $(git describe --tags --abbrev=0)
# If the above doesn't checkout the latest release, this should work:
# git checkout $(git describe --tags `git rev-list --tags --max-count=1`)

Edit the .env file

Copy .env.example to .env and fill in the appropriate values.

cp .env.example .env

Note: .env is optional. If you're planning to run multiple distinct agents, you can pass secrets through the character JSON

Start Eliza

pnpm i
pnpm build
pnpm start

# The project iterates fast, sometimes you need to clean the project if you are coming back to the project
pnpm clean

Interact via Browser

Once the agent is running, you should see the message to run "pnpm start:client" at the end.

Open another terminal, move to the same directory, run the command below, then follow the URL to chat with your agent.

pnpm start:client

Then read the Documentation to learn how to customize your Eliza.

---

Automatically Start Eliza

The start script provides an automated way to set up and run Eliza:

sh scripts/start.sh

For detailed instructions on using the start script, including character management and troubleshooting, see our Start Script Guide.

Note: The start script handles all dependencies, environment setup, and character management automatically.

---

Modify Character

1. Open packages/core/src/defaultCharacter.ts to modify the default character. Uncomment and edit.

2. To load custom characters:

   • Use pnpm start --characters="path/to/your/character.json"
   • Multiple character files can be loaded simultaneously

3. Connect with X (Twitter)

   • change "clients": [] to "clients": ["twitter"] in the character file to connect with X

---

Additional Requirements

You may need to install Sharp. If you see an error when starting up, try installing it with the following command:

pnpm install --include=optional sharp

---

Start Eliza with Gitpod

---

Deploy Eliza in one click

Use Fleek to deploy Eliza in one click. This opens Eliza to non-developers and provides the following options to build your agent:

1. Start with a template
2. Build characterfile from scratch
3. Upload pre-made characterfile

Click here to get started!

---

Community & contact

• GitHub Issues. Best for: bugs you encounter using Eliza, and feature proposals.
• Discord. Best for: sharing your applications and hanging out with the community.

Citation

We now have a paper you can cite for the Eliza OS:

@article{walters2025eliza,
  title={Eliza: A Web3 friendly AI Agent Operating System},
  author={Walters, Shaw and Gao, Sam and Nerd, Shakker and Da, Feng and Williams, Warren and Meng, Ting-Chien and Han, Hunter and He, Frank and Zhang, Allen and Wu, Ming and others},
  journal={arXiv preprint arXiv:2501.06781},
  year={2025}
}

Contributors

Star History