Goose

AI CodingAgentic AI Foundation (originally built by Block) - est. 2024

The free, open-source AI agent you run yourself, as a desktop app or CLI

4.2 / 5Free plan availableVerified review
Try Goose Free →

Start free, upgrade anytime

Reviewed Nov 2026

Editorial take

Goose gives you agentic capabilities close to Claude Code for free: the agent costs nothing and you pay only for the model you connect, whether that is an API key, a local model or an existing Claude, ChatGPT or Gemini subscription. Originally built by Block and now governed by the Linux Foundation's Agentic AI Foundation under Apache 2.0, it is the strongest free alternative to paid coding agents. Note that autonomous mode is the default, so switch to manual or smart approval if you want to review actions first.

What is Goose?

Goose is an open-source AI agent that runs on your own machine. It was originally built by Block (the company behind Square and Cash App), released in early 2025, and contributed to the Linux Foundation's Agentic AI Foundation (AAIF) in December 2025, where it is now developed as a vendor-neutral, community-governed project. It is written in Rust and licensed under Apache 2.0.

Goose ships as a native desktop app for macOS, Linux and Windows, a full CLI, and an API for embedding it in other applications. It can read and edit files, run shell commands, and carry out multi-step tasks such as scaffolding projects, refactoring, running tests and fixing bugs. The project also positions it for non-coding work like research, writing, automation and data analysis. Extensions use the Model Context Protocol (MCP), with 70+ documented extensions, and repeatable workflows can be saved as YAML recipes and shared.

Goose works with dozens of model providers, including Anthropic, OpenAI, Google, Ollama, OpenRouter, Azure and Amazon Bedrock. Instead of API keys, you can use an existing Claude, ChatGPT or Gemini subscription through the Agent Client Protocol (ACP), and the desktop app can run local models in-process without Ollama. The software itself is free; you pay only for the model provider you choose. Approval modes range from fully autonomous (the default) to manual approval and smart approval, which runs low-risk actions automatically and asks before higher-impact ones.

Best for

Developers and technical users who want a free, vendor-neutral agent for coding and other tasks, using models or subscriptions they already have

Key strength

Full agentic capabilities for free, with any model provider

Output Quality0.0Integration0.0Value0.0Ease of Use3.0Free Tier0.0

Score breakdown (out of 5)

What you would use it for

  • →Running multi-file coding tasks autonomously without a paid agent subscription
  • →Choosing any AI model (Claude, GPT, Gemini or a local model) as the backend for agentic tasks
  • →Keeping code on your machine by running the agent entirely locally
  • →Automating repetitive development tasks like schema migrations or scaffolding
  • →Setting up and running tests as part of a complete agentic development session

Pros & Cons

👍 Pros

  • ✓Free and open source - you only pay your model provider
  • ✓Can use an existing Claude, ChatGPT or Gemini subscription instead of per-token API billing
  • ✓Wide model choice, including local models
  • ✓Desktop app means you do not need to work in the terminal
  • ✓Vendor-neutral governance under the Linux Foundation's Agentic AI Foundation
  • ✓Large MCP extension ecosystem

👎 Cons

  • ✗You have to choose and configure a model provider yourself
  • ✗API costs accumulate on large tasks
  • ✗Autonomous mode is the default, so it edits files and runs commands without asking unless you switch modes
  • ✗Built-in local models are slow to load and have small context windows

Key Features

  • ✓ Native desktop app for macOS, Linux and Windows, plus a CLI and an embeddable API
  • ✓ Reads and edits files and runs shell commands
  • ✓ Dozens of model providers, including Anthropic, OpenAI, Google, Ollama, OpenRouter, Azure and Amazon Bedrock
  • ✓ Uses existing Claude, ChatGPT or Gemini subscriptions through ACP
  • ✓ Built-in local model inference in the desktop app
  • ✓ Autonomous, manual approval, smart approval and chat-only modes
  • ✓ 70+ documented MCP extensions, including MCP Apps with interactive UIs
  • ✓ Reusable YAML recipes and parallel subagents
  • ✓ Works as an ACP agent in Zed, JetBrains IDEs and VS Code
  • ✓ Prompt injection detection, sandbox mode and an adversary reviewer
  • ✓ Open source (Apache 2.0 license)

Available on

MacWindowsLinuxCLIAPI

Integrates with

Anthropic Claude (API or Claude Code subscription via ACP)OpenAI (API or ChatGPT Plus/Pro sign-in)Google Gemini (API or Google account sign-in)Ollama, LM Studio and built-in local inferenceOpenRouter, Azure, Amazon Bedrock and GCP Vertex AIGitHub Copilot (as a model provider)Zed, JetBrains IDEs and VS Code (via ACP)GitHub (MCP extension)Shell environment (arbitrary commands)

Goose Pricing

✅ Goose has a free plan - no credit card required to start.

Open Source

Free/forever
  • ✓Desktop app, CLI and embeddable API
  • ✓Bring your own API keys or use an existing Claude, ChatGPT or Gemini subscription
  • ✓Dozens of model providers plus local models
  • ✓70+ documented MCP extensions
  • ✓Recipes and parallel subagents
  • ✓Apache 2.0 license
Get Open Source →

Goose vs Competitors

Best Goose alternatives →Best AI Coding tools →

AI models powering Goose

From the blog

Developer resources

Related Tools

AI Writing4.8

Claude is the best AI assistant for long-document work and research where accuracy matters. Its context window of up to 1M tokens and resistance to sycophancy make it the honest choice when you need a thinking partner rather than an agreeable assistant. Start on the free tier; the Pro plan at $20/month is worth it if you regularly hit the free usage limits, and Max ($100/month) or Ultra ($200/month) suit heavier daily use.

Free + paid plans|Re-verified Nov 2026
AI Writing4.7

ChatGPT is the most capable all-in-one AI assistant available. Its combination of image generation, voice mode, code execution, web browsing, and a large custom GPT ecosystem makes it the default choice for users who want one tool that handles everything. The tradeoff is that it is rarely the deepest tool on any single task.

Free + paid plans|Re-verified Nov 2026
AI Coding4.7

Cursor is the most powerful AI code editor available. Its multi-file editing is ahead of plugin-based alternatives, making large-scale refactors practical in ways GitHub Copilot cannot match. At $30/month for Premium (Grok models included, other models billed at API rates) it is a meaningful upgrade for any developer doing more than small isolated changes.

Free + paid plans|Re-verified Oct 2026
AI Coding4.5

GitHub Copilot is the right starting point for developers adopting AI-assisted coding. It integrates into whatever editor you already use, requires no workflow change, and has a usable free tier. The Pro plan at $10/month is the lowest-cost path to unlimited AI completions, though chat and agent work now draw from a monthly credit allowance; heavy agent users should look at Pro+ ($39) or Max ($100).

Free + paid plans|Re-verified Nov 2026

This page contains affiliate links. We may earn a commission at no extra cost to you. Learn more.