Goose

AI Coding Assistants AI Productivity Tools AI Developer APIs & Platforms AI No-Code / Automation Tools

↳AI Pair Programming & Terminal Agents ↳AI Code Generation & Autocomplete ↳AI Agent Builders ↳AI Code Review & Quality

Free

Block's open-source AI agent — any model, any tool, runs locally on your machine

0(0 votes)

0 views

Released 2025

Visit Website

Gallery

4 items

VIDEO

About Goose

Goose is an open-source, extensible AI coding agent built by Block (the company behind Square, Cash App, and TIDAL) and released under the MIT license in January 2025 — after nine months of internal development and active daily use by 5,000+ Block employees. It was the first open-source AI agent to support MCP (Model Context Protocol) when MCP was announced in November 2024, earning it early recognition as the pioneer of MCP-native agentic development. The community grew from zero to 55,000+ members with 373+ open-source contributors in the first year, accumulating 23,000+ GitHub stars.

Goose goes beyond code suggestions to install dependencies, execute commands, edit files, run tests, browse the web, and complete multi-step engineering and knowledge work tasks end-to-end. It runs locally on the developer's machine via a CLI or a desktop app (macOS, Windows, Linux), pairing LLM reasoning with a modular library of MCP-based tool integrations. Because it is MCP-first by design, Goose is a blank slate: its capabilities are defined entirely by which MCP servers it connects to (GitHub, Google Drive, Slack, Jira, Confluence, databases, and more) rather than hardcoded specialties. It supports any LLM via BYOK — Claude, GPT, Gemini, DeepSeek, Llama, Mistral, Ollama (local models) — with no vendor lock-in.

Jack Dorsey (Block CEO) has publicly endorsed Goose as central to Block's AI strategy, calling for "open by default" as a North Star. Block uses Goose for autonomous code fixes, incident response, migrations, and maintenance workflows — reporting it can automate 90% of code fixes autonomously. The 2026 roadmap focuses on local-first inference (built-in model downloads with no external runtimes), composable UI primitives, and a stable protocol boundary for the client-server-extension architecture.

Key Features

MCP-first architecture — first open-source agent to support MCP; capabilities are defined entirely by pluggable MCP servers (GitHub, Slack, Jira, databases, and more)
Model-agnostic BYOK — connect Claude, GPT, Gemini, DeepSeek, Llama, Mistral, or any local model via Ollama with no vendor lock-in
Autonomous multi-step task execution — installs dependencies, executes shell commands, edits files, runs tests, browses the web, and completes tasks end-to-end
CLI and desktop app — available as a terminal CLI or a native desktop app on macOS, Windows, and Linux with a shared architecture
Container Use integration — MCP server for git-branch-isolated containerized development with easy rollbacks and safer experimentation
Custom extensions and recipes — share reusable agent configurations (recipes) and build extensions for team-specific workflows
Multi-agent orchestration — run multiple Goose instances in parallel or chain agents for complex multi-step pipelines
Local inference support — run fully private, offline agentic workflows with Ollama and open-weight models at zero API cost

Pros

100% free, MIT licensed, fully open-source — no subscriptions, no seats, no vendor lock-in ever
First open-source agent to support MCP (November 2024) — the most MCP-native agent in the ecosystem
Used by 5,000+ Block engineers daily for production engineering tasks — the most battle-tested open-source agent at enterprise scale
Supports local Ollama models for fully private, offline agentic workflows with zero API cost
Jack Dorsey (Block CEO) publicly committed to 'open by default' — strong organizational backing from a major publicly traded company
55,000+ community members and 373+ contributors with active 2026 roadmap — the most active open-source agent community after OpenHands

Cons

No hosted cloud tier — requires local installation and LLM API key management, which can be a barrier for non-technical team members
Desktop app still maturing — feature parity between CLI and desktop app is not complete as of mid-2026
MCP-first design means capabilities depend on available MCP servers — teams with niche tools may need to build custom extensions
Built-in inference (local model downloads without Ollama) was planned for Q1 2026 but still in progress as of mid-2026
Community-supported rather than commercially supported — no SLA, no dedicated enterprise support team, no paid tier for priority issues

Pricing

Goose is completely free and open-source under the MIT license. There are no subscription fees, seat charges, or usage limits imposed by Block. The only cost is LLM inference, which you pay directly to your chosen provider at their published rates (or run for free locally via Ollama and open-weight models).

Installation is via a single install script on macOS/Linux or a desktop app download for Windows. No account creation, no API key from Block, no sign-up required to start. The CLI and desktop app are both free and open-source.

For enterprise teams at Block-scale, Block operates its own internal deployment of Goose with enterprise-specific extensions. No public enterprise pricing exists since the product is community-maintained under MIT license.

Claim Verified Creator Badge

Are you the founder of Goose? Display this listing's verified badge on your website to show your customers that your product has been vetted and listed on AI Central Resources.

HTML Embed Code

<a href="https://www.aicentralresources.com/tool/goose" target="_blank" rel="noopener">
  <img src="https://www.aicentralresources.com/badges/featured-badge-dark.svg" alt="Featured on AICentralResources" width="200" height="54" style="border: none;" />
</a>

* Place this HTML snippet in your website's footer, landing page, or press section. This creates a search-friendly backlink directly to your verification page.

Connect with Goose

X (Twitter)Discord GitHub

Frequently Asked Questions

Goose is an open-source AI agent by Block (Square) that runs on your local machine. Unlike GitHub Copilot or Cline which are primarily IDE plugins, Goose is a standalone CLI and desktop agent that goes beyond code suggestions to autonomously install dependencies, run commands, edit files, run tests, browse the web, and complete end-to-end tasks. Its MCP-first architecture lets you extend it with any tool integration.

Yes. Goose is MIT licensed and completely free. Install it with one script: curl -fsSL https://github.com/block/goose/releases/download/stable/download_cli.sh | sh. You provide your own LLM API keys (Claude, GPT, Gemini, etc.) or use local Ollama models for free inference.

Goose was the first open-source AI agent to support MCP when it was unveiled by Anthropic in November 2024. Because Goose's entire capability set is defined by MCP servers rather than hard-coded tools, it was immediately extensible to any MCP integration without changes to the core agent. The community commonly refers to Goose as the 'MCP client' reference implementation.

Yes. Goose supports any LLM via BYOK: Claude (all models), GPT and o-series (OpenAI), Gemini (Google), DeepSeek, Llama, Mistral, and any other model accessible via API. For fully local, offline agentic workflows with no API cost, Goose integrates with Ollama to run open-weight models directly on your machine.

Goose was built at Block (formerly Square) by Jack Dorsey's team in San Francisco. It was developed internally for about nine months before the MIT-licensed public release in January 2025. Block uses Goose across 5,000+ engineers at Square, Cash App, and TIDAL for production engineering tasks including code fixes, incident response, migrations, and maintenance automation.