The Google Workspace MCP Server, created by Taylor Wilsdon, is a production-ready implementation of the Model Context Protocol (MCP). It is designed to connect AI assistants, such as Claude, Cursor, or Gemini, to Google Workspace tools.

Key features include:

  • Comprehensive Integration: It offers access to Gmail, Drive, Docs, Sheets, Calendar, Slides, Forms, Tasks, Chat, and Contacts.
  • Read-Write Capabilities: This server supports read, write, and delete actions, such as sending emails or creating calendar events.
  • Advanced Authentication: Built on FastMCP, it supports OAuth 2.1 and Google Desktop OAuth clients. This simplifies setup.
  • Multi-User Support: It is designed for both local single-user and multi-user environments. It uses bearer tokens for secure authentication in custom web applications.
  • Tool Tiers: Users can manage API quota usage by launching the server in different “tiers”. 

To get started:

  1. Repository: Access the source code and documentation on the official Google Workspace MCP GitHub repository.
  2. Google Cloud Setup:
    • Create a project in the Google Cloud Console.
    • Enable the necessary APIs (Gmail, Drive, etc.).
    • Create OAuth 2.0 Credentials for a Desktop Application and download the client JSON file.
  3. Installation: The server can be installed via CLI. For local use with Claude Desktop, add the server configuration to your claude_desktop_config.json file.
  4. Simplified Setup: Community forks, like gwmcp, offer a setup wizard based on Wilsdon’s original project.