taylorwilsdon/google_workspace_mcp

What it solves

This project provides a comprehensive bridge between AI assistants and the Google Workspace ecosystem. It solves the problem of limited or fragmented AI integrations with Google services by offering a single, feature-rich server that allows LLMs to read, write, and manage data across multiple Google apps using natural language commands.

How it works

It implements the Model Context Protocol (MCP), allowing it to connect to various AI clients (like Claude Desktop, ChatGPT, or VS Code) via stdio or streamable HTTP. The server uses OAuth 2.1 for secure authentication and provides over 120 tools across 12 major Google services. It supports multiple deployment modes, including stateless container deployment and multi-user authentication for organizational use, and includes a CLI for direct tool execution.

Who it’s for

• AI Power Users: Individuals wanting deep, natural language control over their Google Workspace.
• Developers: Those building AI agents or tools that need to interact with Google APIs securely.
• IT/Security Teams: Organizations looking to host a centralized, secure, and auditable AI gateway to their Workspace data.

Highlights

• Extensive Coverage: Supports 12 services including Gmail, Drive, Calendar, Docs, Sheets, Slides, Forms, Tasks, Contacts, Chat, Custom Search, and Apps Script.
• Flexible Tool Tiers: Offers core, extended, and complete tiers to manage the LLM's context window by limiting the number of active tools.
• Enterprise-Ready Security: Features OAuth 2.1 multi-user support, stateless deployment, and sensitive path blocking to prevent unauthorized file access.
• Broad Client Compatibility: Works with Claude (Desktop, Web, Mobile, Code), ChatGPT, LM Studio, and any other MCP-compliant client.