Frontier Now
Search

Workplace#Email

Google Workspace MCP: Gmail, Calendar, Docs and Sheets in one server

An open-source server covering 12 Google services, from Gmail and Calendar to Docs and Sheets, with tiered tools, a read-only mode and team deployment.

Author
Frontier Now Editorial
Published

Project and installation docs

View project

https://github.com/taylorwilsdon/google_workspace_mcp

Work is spread across Gmail, Calendar, Drive, Docs and Sheets. Asking an agent to “find the quote from last week’s client email, put it in the shared sheet and set up a call” means connecting several services at once. Workspace MCP puts 12 Google services and more than 120 tools behind one server, authorized with your own Google Cloud OAuth client, so data only travels between you and Google’s APIs.

What it does

  • Services: Gmail, Calendar, Drive, Docs, Sheets, Slides, Forms, Tasks, Contacts and Chat, among others. Chat needs a Workspace account and a one-time app setup.
  • Tool tiers: --tool-tier core, extended or complete controls how many tools load; --tools gmail drive calendar picks services; --read-only removes every write.
  • Team deployment: OAuth 2.1 multi-user auth, a stateless container mode and service accounts with domain-wide delegation let an organization host one shared instance.
  • CLI included: workspace-cli lists and calls tools against a running server, handy for scripts.

Who it’s for

  • Individuals and small teams who spend the day triaging email and juggling meetings.
  • IT admins who want one governed Workspace entry point for their company’s agents.

Setup

Create an OAuth client in Google Cloud Console and enable the APIs you need, then:

export GOOGLE_OAUTH_CLIENT_ID="..."
export GOOGLE_OAUTH_CLIENT_SECRET="..."
uvx workspace-mcp --tool-tier core

For Claude Code, start the server in HTTP mode and run claude mcp add --transport http workspace-mcp http://localhost:8000/mcp.

Our take

About 3,300 stars as of 2026-10-06 and the most complete community Google Workspace server. Compared with the Google connectors built into Claude and ChatGPT, it wins on granularity: precise edits to document paragraphs and sheet cells, per-service and read/write permissions, and self-hosting. The cost is setup. You create a Google Cloud project, enable APIs and configure redirect URIs. The quick-start guide puts it at about five minutes, but it is still a hurdle if you have never touched Google Cloud. The project itself flags the biggest risk: emails, docs and calendar invites can carry prompt injections, so decide which write tools to enable and start with --read-only if unsure. Licensed MIT, no CLA.