llms.txt Content
# Google Workspace MCP Server
> The most feature-complete Google Workspace MCP server — 12 services, 100+ tools, OAuth 2.1 multi-user auth, full CLI, and custom connector support. Open source under MIT license.
The Google Workspace MCP Server connects AI assistants to all major Google Workspace services through the Model Context Protocol (MCP). Built with FastMCP for optimal performance, it supports OAuth 2.1 multi-user authentication, streamable HTTP and stdio transports, pluggable storage backends (Memory, Disk, Valkey/Redis), and stateless container deployments. Works with all free Google accounts and Google Workspace plans.
## Quick Install
- Claude Desktop / web / cowork: Use the custom connector flow — Connectors > Add custom connector > paste your MCP server URL
- Claude Code: `claude mcp add --transport http workspacemcp http://localhost:8000/mcp`
- CLI: `uvx workspace-mcp` (requires Python 3.10+ and uv)
- Docker: Containerized deployment available
## Documentation
- [Quick Start Guide](https://workspacemcp.com/quick-start): Step-by-step installation and setup in 5 steps
- [Guides](https://workspacemcp.com/guides): Client-specific setup articles for Workspace MCP
- [Claude Custom Connector Guide](https://workspacemcp.com/guides/claude-connectors): Connect Workspace MCP Cloud to Claude Desktop, Claude on the web, Claude cowork, and scheduled tasks
- [ChatGPT Developer Mode Guide](https://workspacemcp.com/guides/chatgpt-developer-mode): Connect Workspace MCP to ChatGPT using Developer Mode and a remote MCP server
- [Full API Documentation](https://workspacemcp.com/docs): Complete tool reference for all 12 services with parameters
- [Advanced Deployment](https://workspacemcp.com/docs/deployment): Reverse proxy setup, origin validation, credential store backends (GCS/CMEK), and full environment variable reference
- [FAQ & Client Setup](https://workspacemcp.com/welcome/faq): Claude Desktop connector setup, troubleshooting, and common questions
- [Main R