Skip to content
Get packs

Eliza

An open-source TypeScript framework and app stack for building autonomous AI agents.

GitHub Download pr-evidence-11
Stars
19.8k
Forks
5.9k
Licence
MIT
Language
TypeScript
Last update
4 days ago
Type
Framework

What it is

elizaOS is an open-source TypeScript project for running autonomous AI agents. It includes a core runtime, a user-facing Eliza app for web, desktop, and mobile, a command-line tool, cloud services, and first-party plugins that add abilities such as chat, memory, messaging connectors, and browser automation. The README describes it as model-agnostic, so it can use local models or hosted model providers.

  • Run the Eliza app from source with local agent mode using your own model provider keys
  • Start the standalone agent host as an API-only service with bun run start
  • Embed the @elizaos/core runtime directly in your own TypeScript project
  • Build a first-party style plugin that registers actions, providers, or services
  • Run the local Eliza Cloud stack with mocks for testing

Great for

  • Developers who want to build AI agents or plugins in TypeScript
  • People who want a personal AI assistant they can run on their own hardware or with their own provider keys
  • Teams experimenting with local inference using the Eliza-1 on-device path
  • Builders who want a modular plugin system for actions, providers, and services

Not ideal for

  • Beginners who want a simple one-click chat app: the repository setup assumes Bun, Node versions, and command-line work
  • People who need a stable, minimal library: the README describes a large monorepo with many packages and optional cloud services
  • Anyone who wants to submit third-party plugins: the README says third-party plugins and registry listings are no longer accepted
Step-by-step guide: how to use ElizaInstall, first run, examples, tips and FAQs
Eliza alternativesEliza vs LangChainCrewAI vs Eliza

How to get it

  1. 1

    Clone and build from source

    You'll need: Bun (version pinned in package.json) · Node.js (version pinned in package.json) · Git · Python 3.11+ (only for benchmarks)

    Clone the repository

    git clone --filter=blob:none https://github.com/elizaos/eliza.git

    Enter the folder

    cd eliza

    Install dependencies

    bun install

    Start the app

    bun run dev
  2. 2

    Run it

    • Install the Bun and Node versions pinned in package.json
    • Clone the repository and run bun install
    • Run bun run dev to start the app
    • Or run bun run dev:local to use a local agent with your own provider keys without Cloud sign-in
  3. 3

    Get the most out of it

    • Use bun run --cwd <package> <script> to scope commands to one package
    • Read each package's README before making changes, as the repository asks
    • Use bun run verify to run dependency, type, lint, and audit checks
    • Configure local inference or direct model providers to avoid depending on Eliza Cloud
    • Explore the plugin component guide at docs.elizaos.ai to add actions, providers, or services

Downloads

Source code (.zip)develop branch · straight from GitHubRelease pr-evidence-118 files incl. 26518-release-notes-before-after-v2.jpgOfficial website & docselizaresearch.ai/

Strengths

  • Open source under the MIT license, with a public repository and documentation links
  • Model-agnostic design that supports local, direct-provider, or cloud model backends
  • Modular plugin structure with documented component types and a clear package layout

Trade-offs

  • Large, fast-changing monorepo; the README is truncated and some capabilities depend on OS, permissions, and installed plugins
  • Contribution rules are strict and the project does not accept third-party plugins
  • Some features such as live benchmarks may incur model provider costs

How it compares

ProjectStarsTypeLevelUpdatedBest for
eliza (this)19.8kFrameworkadvanced4 days agoDevelopers who want to build AI agents or plugins in TypeScript
langchain149kFrameworkintermediate4 days agoPython developers building LLM-powered applications
crewAI60.4kLibraryintermediate4 days agoPython developers prototyping multi-agent workflows
deer-flow84.9kAI model / agentadvanced4 days agoDevelopers who want to experiment with multi-agent workflows on their own infrastructure
botpress15kDev toolintermediate6 days agoDevelopers who want to write integrations and bots as TypeScript code using the Botpress CLI and SDK
claude-context12.7kDev toolbeginner4 months agoDevelopers working in large codebases where loading everything into an agent's context is impractical

Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.