Connecting

Add Feather to your AI client in under a minute. All clients connect to the same URL and authenticate with your Feather API key.

Claude Desktop

Claude Desktop connects to remote MCP servers via mcp-remote, a small bridge that runs locally. No install step is needed — npx downloads it automatically on first use.

Open or create ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows) and add the feather entry under mcpServers:

{
  "mcpServers": {
    "feather": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.feather.so",
        "--header",
        "Authorization: Bearer fk_live_YOUR_KEY"
      ]
    }
  }
}

Cursor

Open Cursor Settings → MCP and add a new server, or edit ~/.cursor/mcp.json (global) / .cursor/mcp.json (project) directly:

{
  "mcpServers": {
    "feather": {
      "url": "https://mcp.feather.so",
      "headers": {
        "Authorization": "Bearer fk_live_YOUR_KEY"
      }
    }
  }
}

VS Code

Create a .vscode/mcp.json file in your project root:

{
  "servers": {
    "feather": {
      "type": "http",
      "url": "https://mcp.feather.so",
      "headers": {
        "Authorization": "Bearer fk_live_YOUR_KEY"
      }
    }
  }
}

Any other MCP client

Point the client at https://mcp.feather.so using the Streamable HTTP transport and set a custom request header:

Authorization: Bearer fk_live_YOUR_KEY

After connecting

Restart your AI client after saving the config. The Feather tools will appear automatically — you can verify by asking your AI assistant "What Feather tools do you have access to?"