Goose

v1.0.0Coding Agentsstable

an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM

acpaiai-agentsmcp
Share:
45,656
Stars
0
Downloads
0
Weekly
0/5

What is Goose?

Goose is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to open source, extensible ai agent that goes beyond code suggestions - install, execute, edit, and test with any llm

an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM

This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • an open source, extensible AI agent that goes beyond code su

Use Cases

AI agent development
Code execution and testing
Multi-LLM support
aaif-goose

Maintainer

LicenseApache-2.0
Languagerust
Versionv1.0.0
UpdatedMay 22, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx goose

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Goose

Goose is an open-source, extensible AI agent built in Rust that goes beyond code suggestions to autonomously install packages, execute commands, edit files, run tests, and manage full development workflows on your local machine. It supports 15+ AI providers including Anthropic, OpenAI, Google, and Ollama, and integrates with 70+ extensions via the Model Context Protocol. Developers use Goose for complex, multi-step tasks — from setting up development environments to running automated research and data analysis pipelines — without manual handholding at each step.

Prerequisites

  • macOS, Linux, or Windows operating system
  • curl available for CLI installation (or download the desktop app)
  • An API key for at least one supported LLM provider (Anthropic, OpenAI, Google, Ollama, etc.)
  • An MCP client or use Goose's own desktop app or CLI as the agent interface
1

Install the Goose CLI

Run the one-line installer script which downloads the appropriate binary for your platform and adds it to your PATH.

curl -fsSL https://github.com/aaif-goose/goose/releases/download/stable/download_cli.sh | bash
2

Verify the installation

Confirm Goose installed correctly by checking the version and listing available commands.

goose --version
goose --help
3

Configure your LLM provider

Set up your preferred AI provider. Goose supports Anthropic, OpenAI, Google Gemini, Azure, Ollama, and others. Set the appropriate environment variable for your chosen provider.

# For Anthropic Claude:
export ANTHROPIC_API_KEY=sk-ant-your-key-here

# For OpenAI:
export OPENAI_API_KEY=sk-your-key-here

# For local Ollama:
# Start ollama first, then Goose will auto-detect it
4

Start a Goose session

Launch an interactive Goose session in your project directory. Goose will have access to your filesystem, terminal, and any configured MCP extensions.

cd ~/your-project
goose session
5

Add MCP extensions to Goose

Extend Goose's capabilities by adding MCP servers as extensions. This lets Goose use external tools like database connectors, API clients, or specialized search tools.

goose extension add --name my-extension --command "npx my-mcp-server"

Goose Examples

Client configuration

Configure Goose as an MCP server in claude_desktop_config.json, or use Goose's native CLI directly for autonomous agent sessions.

{
  "mcpServers": {
    "goose": {
      "command": "goose",
      "args": ["mcp"],
      "env": {
        "ANTHROPIC_API_KEY": "sk-ant-your-key-here",
        "GOOSE_PROVIDER": "anthropic",
        "GOOSE_MODEL": "claude-3-5-sonnet-20241022"
      }
    }
  }
}

Prompts to try

Use these prompts in a Goose session to leverage its autonomous multi-step agent capabilities.

- "Set up a new Python FastAPI project with a PostgreSQL database, create the project structure, install dependencies, and write a basic health check endpoint"
- "Find all TODO comments in this codebase, create GitHub issues for each one, and generate a summary report"
- "Run the test suite, identify failing tests, attempt to fix them, and run the tests again to confirm the fixes"
- "Research the top 5 Python web scraping libraries, compare their features, and write a recommendation report in markdown"

Troubleshooting Goose

Goose fails to connect to the LLM provider

Verify the API key environment variable is exported in the same shell session where you run Goose. For Anthropic use ANTHROPIC_API_KEY; for OpenAI use OPENAI_API_KEY. Run 'echo $ANTHROPIC_API_KEY' to confirm it is set.

Goose CLI command not found after installation

The installer adds the binary to ~/.local/bin. If this is not in your PATH, add 'export PATH=$HOME/.local/bin:$PATH' to your ~/.bashrc or ~/.zshrc and reload your shell with 'source ~/.zshrc'.

Goose agent loops or gets stuck on a task

Press Ctrl+C to interrupt the current session. Goose saves session history so you can restart and ask it to continue from where it stopped. For complex tasks, break them into smaller subtasks in your prompt to reduce the chance of the agent losing context.

Frequently Asked Questions about Goose

What is Goose?

Goose is a Model Context Protocol (MCP) server that open source, extensible ai agent that goes beyond code suggestions - install, execute, edit, and test with any llm It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Goose?

Follow the installation instructions on the Goose GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Goose?

Goose works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Goose free to use?

Yes, Goose is open source and available under the Apache-2.0 license. You can use it freely in both personal and commercial projects.

Browse More Coding Agents MCP Servers

Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "goose": { "command": "npx", "args": ["-y", "goose"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Goose?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides