Goose

What is Goose?

Goose is an open-source AI agent created by Block, the financial technology company behind Square and Cash App. It was designed to help developers with coding tasks and broader automation, and it shipped with a strong focus on extensibility from day one.

The agent runs in two forms: a command line interface for terminal lovers and a desktop application for a more visual experience. Both connect to models through the Model Context Protocol (MCP), which means you can plug in cloud models, local models and a growing catalog of extensions.

Because Block built it as an open-source project under the Apache 2.0 license, Goose is free to use, modify and extend, and it has attracted contributions from developers who build their own extensions for it.

Key Features

• CLI and desktop app interfaces
• Works with local and cloud models through MCP
• Extension catalog for adding tools and workflows
• Can read and edit code, run commands and manage files
• Recipe-style reusable automation
• Session history you can revisit and share
• Apache 2.0 open-source license
• Active community with regular releases

How to get started

Install Goose using the install script or your package manager, then launch it in your terminal or as a desktop app. It will walk you through choosing a model provider, and from there you can start asking it to work on your projects.

• Install Goose with the official install script or package manager
• Launch the CLI or desktop app
• Choose a model provider and configure access
• Start a session and ask Goose to work on a task
• Browse the extension catalog to add capabilities

Use cases

• Writing and editing code with AI assistance
• Automating repetitive development tasks
• Running commands and managing files through chat
• Building custom extensions for team workflows
• Using local models for privacy-sensitive work
• Documenting projects and generating reports

Pricing and licensing

Goose is free and open source under the Apache 2.0 license, and it is backed by Block. There is no subscription for the tool itself, though you may need your own model API keys depending on which model provider you choose.

The extension ecosystem lets you add capabilities without paying for a proprietary platform, and the open license means companies can integrate Goose into their own tooling freely.

Pros and cons

Goose's strengths are its open license, corporate backing and extensibility. Having Block behind it gives the project resources and longevity, while the Apache 2.0 license keeps it free and community-driven. The MCP-based extension system is genuinely flexible.

The trade-offs are that it is younger than some rivals, so the ecosystem of extensions is still growing, and automation agents in general need supervision on complex tasks. If you want maximum control over the agent's behavior, you may need to build your own extensions.

Alternatives

• Aider: AI pair programming in the terminal
• OpenCode: open-source terminal AI coding agent
• Cline: autonomous coding agent inside VS Code
• Continue: open-source AI assistant for VS Code and JetBrains
• Cursor: AI-native graphical editor