Overview
goose is a general-purpose, open-source AI agent that runs natively on your machine — available as a desktop app for macOS, Linux, and Windows, a full-featured CLI, and an embeddable API. Built in Rust for performance and portability, it goes well beyond code: use it for research, writing, automation, data analysis, or any task you need to get done.
Developers love goose because it is vendor-neutral and built on open standards. It is one of the earliest and deepest adopters of the Model Context Protocol (MCP), with 70+ documented extensions for databases, APIs, browsers, GitHub, Google Drive, and more. It also speaks the Agent Client Protocol (ACP), working as an ACP server from Zed, JetBrains, or VS Code, and can use ACP agents like Claude Code and Codex as providers.
Now part of the Agentic AI Foundation (AAIF) at the Linux Foundation, goose is community-governed and open for the long term, with an active community of developers, contributors, and users building extensions, sharing recipes, and pushing the boundaries of local AI agents.
Key Features
- Multi-surface access: Native desktop app (macOS, Linux, Windows), a full CLI for terminal workflows, and an API to embed goose anywhere.
- Extensible via MCP: Connect to 70+ extensions — databases, APIs, browsers, GitHub, Google Drive — via the Model Context Protocol open standard, and add skills or build your own extension.
- Any LLM, including your subscriptions: Works with 15+ providers (Anthropic, OpenAI, Google, Ollama, OpenRouter, Azure, Bedrock, and more) using API keys or existing Claude, ChatGPT, or Gemini subscriptions via ACP.
- Recipes: Capture workflows as portable YAML configs to share with your team, run in CI, and compose with instructions, extensions, parameters, and subrecipes.
- MCP Apps: Extensions can render interactive UIs — buttons, forms, visualizations — directly inside goose Desktop.
- Subagents: Spawn independent subagents to handle tasks in parallel (code review, research, file processing) while
