Gmail MCP Server
Enables AI assistants to search, read, and download attachments from Gmail using natural language, with read-only access.
README
Gmail MCP Server
An MCP (Model Context Protocol) server that provides Gmail integration for AI assistants. Search emails, read messages, and download attachments directly from your Gmail account.
Features
- Search emails using Gmail's powerful search syntax (
from:,subject:,has:attachment, etc.) - Read full email content including body and attachment metadata
- Download attachments to your local filesystem
Quick Start
1. Set up Google Cloud credentials
- Go to Google Cloud Console
- Create a new project (or select existing)
- Enable the Gmail API
- Go to APIs & Services > Credentials
- Click Create Credentials > OAuth client ID
- Select Desktop app as application type
- Download the JSON file and save it:
mkdir -p ~/.gmail-mcp
mv ~/Downloads/client_secret_*.json ~/.gmail-mcp/gcp-oauth.keys.json
2. First-time authorization
Run once to complete OAuth (opens browser):
uvx mcp-gmail-reader
Backup & Recovery
After a fresh OS install or losing your machine, here's what to restore:
| File | Where | Backup? | Recovery |
|---|---|---|---|
~/.gmail-mcp/gcp-oauth.keys.json |
Local only | YES — password manager | Cannot regenerate without recreating GCP project |
~/.gmail-mcp/token.json |
Local only | No | Regenerated by running uvx mcp-gmail-reader once (browser flow) |
~/Documents/gmail-mcp/ (this repo) |
Git | No need — git clone |
git clone https://github.com/Maheidem/gmail-mcp.git |
| MCP registration | ~/.claude/settings.json |
No | Re-add the JSON snippet from "Claude Code (CLI)" section above |
Recovery checklist (5 minutes after a fresh machine)
# 1. Clone the source
git clone https://github.com/Maheidem/gmail-mcp.git ~/Documents/gmail-mcp
# 2. Restore credentials from your password manager
mkdir -p ~/.gmail-mcp
# (paste gcp-oauth.keys.json into ~/.gmail-mcp/)
# 3. Install + first-run authorization (opens browser)
uvx mcp-gmail-reader
# 4. Re-register in Claude Code (see "Claude Code (CLI)" section below)
The OAuth credentials JSON is the only thing that cannot be recovered without doing the full Google Cloud Console setup again. Back it up.
Setup by AI Tool
Claude Desktop
Add to your config file:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"gmail": {
"command": "uvx",
"args": ["mcp-gmail-reader"]
}
}
}
Claude Code (CLI)
Add to your MCP settings (~/.claude/settings.json or project .mcp.json):
{
"mcpServers": {
"gmail": {
"command": "uvx",
"args": ["mcp-gmail-reader"]
}
}
}
Cursor
Add to Cursor's MCP config (Settings > MCP Servers):
{
"gmail": {
"command": "uvx",
"args": ["mcp-gmail-reader"]
}
}
Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"gmail": {
"command": "uvx",
"args": ["mcp-gmail-reader"]
}
}
}
VS Code + Continue
Add to Continue's config (~/.continue/config.json):
{
"experimental": {
"modelContextProtocolServers": [
{
"transport": {
"type": "stdio",
"command": "uvx",
"args": ["mcp-gmail-reader"]
}
}
]
}
}
Zed
Add to Zed's settings (~/.config/zed/settings.json):
{
"context_servers": {
"gmail": {
"command": {
"path": "uvx",
"args": ["mcp-gmail-reader"]
}
}
}
}
Custom credentials path
If your credentials are in a different location:
{
"mcpServers": {
"gmail": {
"command": "uvx",
"args": ["mcp-gmail-reader"],
"env": {
"GMAIL_CREDENTIALS_PATH": "/path/to/your/credentials.json"
}
}
}
}
Available Tools
search_emails
Search Gmail using Gmail search syntax.
Query examples:
from:example@gmail.com- emails from specific sendersubject:invoice- emails with subject containing "invoice"has:attachment- emails with attachmentsafter:2024/01/01- emails after dateis:unread- unread emailslabel:important- emails with specific labelin:inbox- emails in inbox
get_email
Get full email content by message ID (returned from search_emails).
Returns: id, from, to, subject, date, body, attachments list
download_attachment
Download email attachment to specified path.
Parameters: message_id, attachment_id, save_path
Development
git clone https://github.com/Maheidem/gmail-mcp.git
cd gmail-mcp
uv sync
# Run server
uv run python -m gmail_mcp
# Run with MCP inspector
uv run mcp dev src/gmail_mcp/server.py
Security Notes
- Only
gmail.readonlyscope is used - this server cannot send or modify emails - OAuth tokens are stored locally in
~/.gmail-mcp/token.json - Never commit credentials or tokens to version control
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.