Overview
ElizaOS is an open-source "operating system" for autonomous AI agents—an integrated platform that combines a modular backend, a React web UI, a command-line interface (CLI), and an extensible plugin ecosystem to help developers build, run, and manage multi-agent AI applications.
Key Capabilities
- Model-agnostic: Integrates with multiple model providers and runtimes, enabling use of OpenAI-style APIs, Anthropic, Gemini, local Llama-style models, and others.
- Multi-agent architecture: Designed to create, orchestrate, and manage groups of specialized agents that can communicate and collaborate to accomplish tasks.
- Rich connectivity: Built-in connectors for popular chat platforms (Discord, Telegram, Farcaster, etc.) and other integrations to deploy agents into real-world channels.
- Document ingestion & RAG: Tools for indexing documents and enabling agents to retrieve and answer from private data sources.
- Extensible plugin system: Developers can add custom functionality via plugins (database adapters, connectors, tool integrations, etc.).
- Developer ergonomics: A modern web dashboard for realtime management, a CLI for quick project scaffolding and local development, and example standalone usage for embedding agents into applications.
Architecture & Components
ElizaOS is maintained as a monorepo containing multiple packages:
- @elizaos/server — core backend (Express.js) that runs agents and exposes APIs.
- @elizaos/client — React-based web UI for managing agents, conversations, and deployments.
- @elizaos/cli — command-line tool to scaffold projects, run agents, and manage environments.
- plugin packages — modular plugins for database integration (Postgres, PGLite), bootstrap/event handling, and other extensions.
- examples — standalone usage samples (interactive chat, basic processing) to embed agents directly into apps.
Getting Started (high level)
-
Install the CLI (requires Node.js and bun):
bun install -g @elizaos/cli
-
Create a new agent project:
elizaos create my-first-agent
-
Configure API keys and environment variables, then start:
elizaos start
The platform supports quick local development, a production-ready server API, and an interactive web dashboard at http://localhost:3000 when running locally.
Use Cases
- Conversational agents (chatbots) and assistant bots for teams.
- Autonomous business process automation where multiple agents coordinate tasks.
- Game NPCs with complex, multi-agent behaviors.
- Knowledge agents that answer from ingested documents (RAG).
Community, License & Citation
ElizaOS is MIT-licensed and accepts community contributions via GitHub. Project documentation is available at its docs site, and the repository includes contribution guidelines, CI status, and badges. The project also provides a citation (arXiv preprint referenced in the repo) for academic use.
Who should use it
Developers and teams who need a full-stack, extensible platform to prototype and deploy multi-agent AI systems quickly—especially when they want vendor flexibility for model providers and a plugin-based architecture for production integration.
