elizaOS

Overview

elizaOS is an open-source TypeScript framework and product stack for autonomous AI agents. The monorepo includes the core runtime, the Eliza app for web, desktop, and mobile, the CLI, cloud services, native bridges, first-party plugins, and Linux and Android distribution tooling.

Key Features

• Monorepo stack includes the core runtime, Eliza app, CLI, cloud services, native bridges, first-party plugins, and Linux and Android distribution tooling.
• The Eliza app supports chat, voice, memory, knowledge, document workflows, messaging and workspace connectors, personal-assistant domains, browser and desktop automation, and scheduled workflows.
• Native device bridges cover camera, phone, messages, contacts, and location.
• Non-custodial EVM and Solana wallet operations include approval boundaries.
• A local inference path provides Eliza-1 on-device models and can route capabilities to local, direct-provider, or Eliza Cloud backends.
• Plugins can register actions, providers, evaluators, services, model handlers, routes, events, tests, and app views.

Use Cases

• Build and run personal AI agents that chat, remember context, manage calendars and reminders, and automate browser or desktop tasks.
• Develop plugins using the runtime and plugin contracts to add actions, providers, services, routes, events, tests, or app views.
• Run Eliza from source in a local-first setup with your own provider keys, without Cloud sign-in.
• Evaluate providers and models with the Python benchmark orchestrator.

Getting Started

• Install the Bun and Node versions pinned in package.json, clone the repository, run bun install, then bun run dev.
• Use bun run dev:local for the app and API with the runtime chooser and your own provider keys; Cloud sign-in is not required.
• Use bun run start for the standalone agent host, which provides the API only and no UI.
• For benchmarks, use Python 3.11+ and the suite dependencies, then run the benchmark orchestrator list and run commands from the repository root.

Deployment & Requirements

• Install the Bun and Node versions pinned in package.json.
• Capability availability depends on the operating system, installed plugins, granted permissions, and configured model or service providers.
• Benchmarks require Python 3.11+ and the dependencies required by the selected suite.
• Live benchmarks require the selected provider's credentials and may incur usage costs.
• Local inference uses hardware detection and model routing to select supported backends; eligible operations can run offline after required assets are downloaded.

Before You Adopt

• License: MIT. Review its terms before using, modifying, or distributing the project.
• Live benchmark runs depend on provider credentials and may add usage costs.
• The repository is a large monorepo with many moving parts, so adopters should expect integration complexity.
• Contribution scope excludes third-party plugins and registry items.