Open-source SDK for building and running AI agents in Python and TypeScript, with tools, MCP, memory, and model portability.
elizaOS/eliza
An open-source TypeScript framework and app stack for building autonomous AI agents, with a core runtime, plugins, and a local-first Eliza app.
- Stars
- 19,634
- Forks
- 5,814
- Open issues
- 233
- Last commit
- 2 hours ago
What is it?
elizaOS is a monorepo for building and running autonomous AI agents. It includes a core runtime (@elizaos/core) that defines agent state, memory, authorization, and a plugin contract, along with a standalone agent host, a shared React UI, and a CLI.
Plugins are the main extension point. A plugin can register actions, providers, evaluators, services, model handlers, routes, and app views. The repository ships first-party plugins for model providers, messaging connectors, domain features, and device bridges, and it includes a local inference path (Eliza-1) based on Gemma 4 text tiers and local embedding, speech, and vision assets.
The Eliza app targets web, desktop, and mobile, and Eliza Cloud is an optional hosted service for accounts, model routing, and deployment. Developers can run the stack from source with Bun, or import @elizaos/core to embed the runtime directly.
Key features
- 01Core runtime with agent state, memory, and plugin contracts
- 02Plugin system for actions, providers, evaluators, services, and routes
- 03Local inference path with Eliza-1 model tiers and device-based routing
- 04Messaging and workspace connectors, plus browser and desktop automation
- 05Non-custodial EVM and Solana wallet operations with approval steps
- 06Model-agnostic: local, direct-provider, or Eliza Cloud backends
Use cases
- Building a custom autonomous agent that connects to chat platforms like Discord, Slack, or Telegram
- Writing a first-party-style plugin that adds model handlers or actions to an agent
- Running a personal assistant with local models on supported hardware
- Embedding the agent runtime in an existing TypeScript application
Quick start
- 1
Install the Bun and Node versions pinned in package.json
- 2
Clone the repository
git clone --filter=blob:none https://github.com/elizaos/eliza.git - 3
Enter the folder and install dependencies
cd eliza && bun install - 4
Start the app in development mode
bun run dev
Things to know
- Running from source needs specific Bun and Node versions and sets up submodules and native build prerequisites.
- Local inference needs supported hardware and downloads model assets before offline use.
- Some capabilities depend on model provider keys or optional Eliza Cloud services.
Topics
FAQ
- What is eliza?
- An open-source TypeScript framework and app stack for building autonomous AI agents, with a core runtime, plugins, and a local-first Eliza app. elizaOS is a monorepo for building and running autonomous AI agents. It includes a core runtime (@elizaos/core) that defines agent state, memory, authorization, and a plugin contract, along with a standalone agent host, a shared React UI, and a CLI.
- Is eliza free to use? What license does it have?
- eliza is open source under the MIT license. Check the license terms before using it in commercial products.
- Is eliza still maintained?
- Actively maintained. The last commit to eliza was on Oct 21, 2026.
- What can I use eliza for?
- Building a custom autonomous agent that connects to chat platforms like Discord, Slack, or Telegram Writing a first-party-style plugin that adds model handlers or actions to an agent Running a personal assistant with local models on supported hardware Embedding the agent runtime in an existing TypeScript application
- How do I install eliza?
- Install the Bun and Node versions pinned in package.json Clone the repository: git clone --filter=blob:none https://github.com/elizaos/eliza.git Enter the folder and install dependencies: cd eliza && bun install Start the app in development mode: bun run dev
- Where can I get eliza?
- The source code and installation instructions are on GitHub: https://github.com/elizaOS/eliza
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
LobeHub is a platform for organizing AI agents into always-on teams that can be hired, scheduled, and reviewed from one place.
Self-hosted personal AI agent framework in Python with a browser WebUI, terminal client, chat app channels, tools, memory and MCP support.
A JavaScript library that embeds a GUI agent in any web page, letting users control the interface with natural language.
Code-first Python framework from Google for building, evaluating, and deploying AI agents, optimized for Gemini but model-agnostic.
Open-source platform for building AI-powered chat bots that connect LLMs to Slack, Discord, Telegram, WeChat, Lark, DingTalk and more.