Goose
Goose is an open-source AI agent automating engineering tasks via desktop, CLI, and API.
About
Goose is an open-source AI agent developed by Block (now part of the Linux Foundation) to automate engineering tasks through a desktop application, command-line interface (CLI), and API. It integrates with over 70 extensions via the Model Context Protocol (MCP), enabling seamless connections to tools like GitHub, Google Drive, databases, and custom APIs. The agent is compatible with more than 15 large language model (LLM) providers, including Anthropic, OpenAI, Google, Mistral, Ollama, OpenRouter, Azure, and Bedrock, allowing users to choose models based on task requirements, cost considerations, or privacy needs. Built in Rust, Goose offers fast startup times, low-latency tool calls, and efficient resource usage, making it suitable for long-running autonomous sessions. Its architecture comprises three main components: the interface (desktop app or CLI), the agent (core logic managing interactions), and connected extensions (tools providing specific capabilities). This design allows Goose to perform tasks such as code generation, refactoring, and complex workflow automation. Initially developed as an internal tool at Block, Goose was open-sourced in 2025 and donated to the Agentic AI Foundation (AAIF) at the Linux Foundation in April 2026, ensuring long-term community governance and development. The project has gained significant traction, with over 45,000 stars on GitHub and contributions from more than 500 developers. Its flexibility and extensibility have made it a valuable resource for developers seeking a local, LLM-agnostic AI agent capable of integrating with various tools and services. Goose's open-source nature and active community support position it as a prominent solution in the AI agent landscape, offering a robust alternative to proprietary coding assistants and fostering innovation in AI-driven development workflows.
◆ Why it matters
Goose is a significant contributor to the AI economy by providing an open-source, LLM-agnostic AI agent that operates locally on users' machines. Its compatibility with over 15 LLM providers and more than 70 extensions via the Model Context Protocol (MCP) enables seamless integration with a wide array of tools and services. The agent's architecture, built in Rust, ensures high performance and efficiency, making it suitable for complex, long-running tasks. By open-sourcing Goose and donating it to the Agentic AI Foundation (AAIF) at the Linux Foundation, Block has fostered a collaborative development environment, leading to substantial community engagement and rapid growth. This strategic move has positioned Goose as a versatile and accessible solution for developers, promoting innovation and flexibility in AI-driven development workflows. Its ability to integrate with various LLMs and tools without vendor-specific dependencies addresses a critical need in the AI agent market, offering a compelling alternative to proprietary solutions and contributing to the diversification and democratization of AI technologies.
Notable for
Goose is notable for being an open-source, LLM-agnostic AI agent that automates engineering tasks across multiple platforms.