A self-hosted proactive AI agent template built on OpenBot that sorts your mail and calendar, works in its own browser, and asks before acting.
taylorwilsdon/google_workspace_mcp
An MCP server and CLI that gives AI assistants tools to control Gmail, Calendar, Drive, Docs, Sheets, Slides, Forms, Tasks, Contacts, Chat and more.
- Stars
- 3,483
- Forks
- 1,097
- Open issues
- 100
- Last commit
- 2 days ago
What is it?
Workspace MCP is a Model Context Protocol server that exposes Google Workspace services as tools for AI assistants and MCP-compatible clients. It covers Gmail, Google Calendar, Drive, Docs, Sheets, Slides, Forms, Tasks, Contacts, Chat, Custom Search and Apps Script, with more than 120 tools in total. Tools are grouped into core, extended and complete tiers so clients load only the set they need, and individual services or tools can be enabled or disabled.
The server runs locally over stdio for older clients or remotely over streamable HTTP. It supports OAuth 2.1 with multi-user authentication, a read-only mode, and a stateless mode for containers that cannot write to disk. Deployment options include Docker, reverse-proxy setups, external OAuth providers, service accounts with domain-wide delegation, and optional OpenTelemetry tracing. A bundled `workspace-cli` tool lists and calls tools from the terminal with cached OAuth tokens.
The project is written in Python, installs via uvx or PyPI, and is MIT licensed.
Key features
- 01Covers 12 Google Workspace services with more than 120 tools behind one MCP server.
- 02Supports OAuth 2.1 multi-user authentication and streamable HTTP transport.
- 03Offers core, extended and complete tool tiers, plus a read-only mode.
- 04Includes a stateless container mode and external OAuth provider support.
- 05Ships a CLI for listing and calling tools from the terminal.
Use cases
- Letting Claude or ChatGPT search, summarize and draft replies to Gmail messages.
- Creating and editing Google Docs or Sheets from natural-language prompts inside an AI client.
- Hosting a shared Workspace MCP endpoint for an organization with per-user OAuth.
- Automating calendar, task or form workflows from scripts through the CLI.
Quick start
- 1
Create an OAuth client in Google Cloud Console and enable the Google APIs for the services you plan to use
- 2
Set the GOOGLE_OAUTH_CLIENT_ID and GOOGLE_OAUTH_CLIENT_SECRET environment variables
- 3
Launch the server with the core tool tier
uvx workspace-mcp --tool-tier core
Things to know
- Requires you to create your own Google Cloud OAuth client and enable the APIs before it works.
- Google Chat needs a one-time Chat app configuration and a Google Workspace account.
- OAuth 2.1 mode requires HTTP transport rather than stdio.
In collections
Topics
FAQ
- What is google_workspace_mcp?
- An MCP server and CLI that gives AI assistants tools to control Gmail, Calendar, Drive, Docs, Sheets, Slides, Forms, Tasks, Contacts, Chat and more. Workspace MCP is a Model Context Protocol server that exposes Google Workspace services as tools for AI assistants and MCP-compatible clients. It covers Gmail, Google Calendar, Drive, Docs, Sheets, Slides, Forms, Tasks, Contacts, Chat, Custom Search and Apps Script, with more than 120 tools in total. Tools are grouped into core, extended and complete tiers so clients load only the set they need, and individual services or tools can be enabled or disabled.
- Is google_workspace_mcp free to use? What license does it have?
- google_workspace_mcp is open source under the MIT license. Check the license terms before using it in commercial products.
- Is google_workspace_mcp still maintained?
- Actively maintained. The last commit to google_workspace_mcp was on Nov 5, 2026.
- What can I use google_workspace_mcp for?
- Letting Claude or ChatGPT search, summarize and draft replies to Gmail messages. Creating and editing Google Docs or Sheets from natural-language prompts inside an AI client. Hosting a shared Workspace MCP endpoint for an organization with per-user OAuth. Automating calendar, task or form workflows from scripts through the CLI.
- How do I install google_workspace_mcp?
- Create an OAuth client in Google Cloud Console and enable the Google APIs for the services you plan to use Set the GOOGLE_OAUTH_CLIENT_ID and GOOGLE_OAUTH_CLIENT_SECRET environment variables Launch the server with the core tool tier: uvx workspace-mcp --tool-tier core
- Where can I get google_workspace_mcp?
- The source code and installation instructions are on GitHub: https://github.com/taylorwilsdon/google_workspace_mcp
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
MCP server for AI coding agents that keeps raw tool output out of the context window and persists session state across compaction.
A Claude Code, OpenCode and Codex skill that runs goal-driven modify-verify-keep/discard loops, inspired by Karpathy's autoresearch.
An email service for autonomous AI agents that lets them register, read, send, and manage their own @atomicmail.ai inbox via JMAP.
An MCP server and CLI that connect your Zotero research library to Claude, ChatGPT, Cursor and other AI assistants for search, reading and annotation.
A self-hosted MCP server that lets ChatGPT, Claude and other MCP hosts read, edit and run code in your local projects.