Goose

Goose is a general-purpose open-source AI agent originally developed by Block (formerly Square) and now part of the Agentic AI Foundation (AAIF) at the Linux Foundation. It runs on your machine as a native desktop app for macOS, Linux, and Windows, and also provides a full CLI and API for terminal workflows and programmatic integration.

Unlike coding-specific agents, Goose is designed as a general-purpose assistant that helps with code development, research, writing, automation, data analysis, and more. Built in Rust for performance and portability, it is licensed under Apache 2.0 and hosted on GitHub at github.com/block/goose.

Desktop App:

Download the native desktop app from goose-docs.ai for macOS, Linux, and Windows. The desktop app provides a graphical interface for managing sessions, configuring extensions, and interacting with the agent.

CLI Installation:

curl -fsSL https://github.com/aaif-goose/goose/releases/download/stable/download_cli.sh | bash

Goose supports 15+ LLM providers, giving you flexibility to choose the models that work best for your tasks:

Provider	Connection Method
Anthropic (Claude)	API key or existing Claude subscription via ACP
OpenAI (GPT)	API key or existing ChatGPT subscription via ACP
Google (Gemini)	API key or existing Gemini subscription via ACP
OpenRouter	API key for unified model access
Ollama	Local model serving
Azure, Bedrock	Enterprise cloud deployment

Goose connects to 70+ extensions through the Model Context Protocol (MCP), an open standard for AI agent tool integration. This ecosystem allows Goose to interact with a wide range of services, databases, APIs, and development tools.

MCP extensions expand Goose's capabilities to include file system operations, database queries, web browsing, API integrations, and custom tool execution. The MCP ecosystem is community-driven and continues to grow as developers publish new extensions.

Goose's native desktop app provides a polished graphical interface for interacting with the AI agent. It supports multi-turn conversations, file attachments, session management, and visual configuration of providers and extensions. The desktop app is built with cross-platform frameworks and feels native on all major operating systems.

• Session Management — Save, resume, and organize conversation sessions
• Extension Browser — Discover and install MCP extensions from the built-in catalog
• Provider Configuration — Visual setup for API keys, models, and provider parameters
• File Attachments — Share files, images, and code with the agent directly

For terminal enthusiasts and automation workflows, Goose provides a full-featured CLI that can be used in scripts and CI/CD pipelines. The API mode allows embedding Goose into other applications, enabling programmatic access to the agent's capabilities.

Goose also supports ACP (Agent Communication Protocol) for interoperability with other AI agents and frameworks. This allows Goose to participate in multi-agent workflows and act as a provider for other agent systems through its ACP interface.

Goose is governed by the Agentic AI Foundation (AAIF) at the Linux Foundation, ensuring open governance and community-driven development:

• Linux Foundation — Neutral governance through AAIF ensures long-term project sustainability
• Custom Distributions — Build your own Goose distro with preconfigured providers, extensions, and branding
• Discord Community — Active community at discord.gg/goose-oss
• GitHub — Open-source development at github.com/block/goose
• Documentation — Comprehensive guides and tutorials at goose-docs.ai