The AI agent that migrated to the Linux Foundation

What it does

goose is a general-purpose AI agent that runs locally on your machine through a native desktop app, a CLI, or an embeddable API. It works with more than 15 LLM providers — Anthropic, OpenAI, Google, Ollama, Azure, Bedrock, and others — and handles tasks that go beyond code suggestions: writing, editing, executing, testing, research, and automation. The project is built in Rust and recently moved from Block to the Agentic AI Foundation at the Linux Foundation.

The interesting bit

Instead of treating extensibility as a plugin store, goose leans on open standards. It connects to 70-plus extensions via the Model Context Protocol, and it supports ACP so you can reuse existing Claude, ChatGPT, or Gemini subscriptions rather than buying fresh API keys. Teams can also build custom distributions — preconfigured bundles of providers, extensions, and branding — without maintaining a hard fork.

Key highlights

• Three interfaces: native desktop app, CLI, and API
• 15-plus LLM providers supported, plus the ability to use existing subscriptions via ACP
• 70-plus extensions through the Model Context Protocol open standard
• Governed by the Agentic AI Foundation (AAIF) at the Linux Foundation
• Custom distributions let you ship a branded, preconfigured version of the agent

Caveats

• The repository recently migrated from block/goose; the README warns that some links and references are still being updated during the transition.

Verdict

Evaluate it if you want an extensible, locally running agent with open governance. Pass if you prefer a managed, cloud-only assistant — goose is built to live on your hardware.