elizaOS/eliza

Open source agentic operating system

What it solves

elizaOS provides a comprehensive framework and product stack for creating and deploying autonomous AI agents. It bridges the gap between a core AI runtime and actual usable products by providing the necessary infrastructure for agents to interact with the real world, including device hardware, third-party apps, and cloud services.

How it works

The system is built as a TypeScript monorepo consisting of several key layers:

  • The Framework: A model-agnostic core (@elizaos/core) that handles memory, state, and plugin contracts, allowing developers to extend agent capabilities via plugins.
  • The App Stack: A user-facing application layer (@elizaos/app and @elizaos/ui) that brings the agents to web, desktop, and mobile platforms.
  • Local Inference: A dedicated plugin that enables on-device execution of text, speech, vision, and image models (using Gemma 4 tiers), allowing agents to function without an internet connection when hardware permits.
  • Cloud Integration: An optional cloud layer for authentication, hosted routing, and cross-device synchronization.
  • OS Distributions: Tooling to package the entire stack into bootable Linux and Android (AOSP) distributions.

Who it’s for

It is designed for developers building autonomous agents, as well as users who want a complete "agentic operating system" experience on their devices.

Highlights

  • Extensible Plugin System: Supports plugins for actions, providers, evaluators, and device bridges.
  • Broad Integration: Includes connectors for messaging, workspaces, calendars, and non-custodial crypto wallets (EVM and Solana).
  • Hybrid Execution: Can route tasks between local on-device inference, direct provider APIs, or cloud backends.
  • Device Automation: Capable of browser and desktop automation, as well as accessing native device features like camera, phone, and location.

Related

  • Project
    framerslab/agentos

    AgentOS is a TypeScript framework for building LLM‑driven AI agents with persistent, neuroscience‑inspired memory, optional HEXACO personality, runtime tool‑forging, and full‑runtime “Generalized Mind Instances.” It supports 13 LLM providers, multi‑agent orchestration, RAG pipelines, guardrails, voice/telephony, and a large extension ecosystem, and it reports top open‑source memory benchmark scores.

  • Project
    luoguoxiong/aipack

    aipack is a TypeScript‑first, lightweight AI‑agent framework. It includes its own runtime, a Tapable‑based plugin system, and built‑in adapters for 13+ LLM providers. Features include streaming vs. batch execution, persistent session storage, permission‑based tool approval, hybrid BM25 + vector long‑term memory, five‑stage context compression, multi‑agent orchestration (graph, pipeline, router, debate, map‑reduce), two‑way MCP integration, full observability (metrics, dashboard), and an offline evaluation suite. The project ships a CLI assistant, a library for embedding in apps, and example apps demonstrating real‑world use cases.

  • Project
    ag2ai/ag2

    An open-source AgentOS framework for building AI agents and orchestrating their cooperation through a hub-and-channel network system.

  • Project
    BuilderIO/agent-native

    A TypeScript framework for building autonomous agents that share a unified action layer, database, and state with a purpose-built user interface.

  • Project
    holaboss-ai/holaOS

    An agentic workspace that integrates AI agents side-by-side with real applications and chat tools, featuring local-first memory and shared tool access.