gmail-mcp
Local stdio MCP server enabling Claude to send Gmail, search emails with attachment metadata, and download attachments to Google Drive.
README
gmail-mcp
A local stdio MCP server that gives your local Claude clients these Gmail tools:
| Tool | What it does |
|---|---|
send_email |
Send mail from your Gmail/Workspace address (plain or HTML, cc/bcc). Every send is logged. |
search_emails |
Search Gmail (has:attachment from:bob newer_than:7d…) and return message summaries including attachment metadata. |
download_attachment |
Pull an attachment from a message and upload it to your Google Drive, returning the Drive link. |
download_attachment_local |
Pull an attachment from a message and save it to a local directory (dest_dir, expanded and created if absent), returning the saved file path. |
Design
- Local stdio — no public endpoint, no OAuth-for-callers, no server to host. Runs as a subprocess of your Claude client.
- Single Google account for both Gmail and Drive.
- Least-privilege scopes —
gmail.send,gmail.readonly,drive.file(the app only ever sees Drive files it created). - Credentials in AWS SSM — the refresh token lives in an SSM SecureString and is fetched at
launch; nothing is written to disk. Requires AWS credentials at runtime (e.g. an
AWS_PROFILE). - Attachments upload into a Drive folder named
Gmail Attachmentsby default (created on first use; override per call with thedrive_folderargument).
Prerequisites
uv(manages Python 3.12 + deps).- AWS credentials with read/write access to the SSM parameter (any standard method —
AWS_PROFILE, env vars, instance role). A region from your AWS config, envAWS_REGION, orGMAIL_MCP_AWS_REGION. - A Google Cloud project with the Gmail API and Google Drive API enabled (Console → APIs & Services → Library).
- A Google OAuth client of type "Desktop app" — download its
client_secret.json(Console → APIs & Services → Credentials).
One-time setup: mint the token
Runs a browser consent for the three scopes and stores the credentials in SSM
(default param /gmail-mcp/authorized-user-json).
cd /path/to/gmail-mcp
# Using a downloaded Desktop client_secret.json:
AWS_PROFILE=<your-aws-profile> \
uv run python scripts/mint_token.py --client-secrets-file /path/to/client_secret.json
# ...or reuse an existing Google token already in SSM (its client_id/secret):
AWS_PROFILE=<your-aws-profile> \
uv run python scripts/mint_token.py --client-secrets-ssm-param /path/to/existing-token
A browser window opens — approve access for the intended account. On success the token is written to SSM and the server can run.
Register with your Claude client(s)
Replace /path/to/gmail-mcp and <your-aws-profile> with your values.
Claude Code
claude mcp add gmail-mcp -s user \
-e AWS_PROFILE=<your-aws-profile> \
-- uv run --directory /path/to/gmail-mcp gmail-mcp
Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):
{
"mcpServers": {
"gmail-mcp": {
"command": "uv",
"args": ["run", "--directory", "/path/to/gmail-mcp", "gmail-mcp"],
"env": { "AWS_PROFILE": "<your-aws-profile>" }
}
}
}
Restart Claude Desktop after editing. The three tools appear under the gmail-mcp server.
Usage examples (things to ask Claude)
- "Email alice@example.com with the subject 'Lunch?' and body 'Free tomorrow at 1?'"
- "Search my Gmail for emails with attachments from the last week."
- "Download the PDF attached to that invoice email into my Drive."
Configuration
| Variable | Default | Purpose |
|---|---|---|
AWS_PROFILE (or any AWS cred) |
— | AWS credentials used to read the token from SSM |
GMAIL_MCP_AWS_REGION |
standard AWS config | Region of the SSM parameter |
GMAIL_MCP_SSM_PARAM |
/gmail-mcp/authorized-user-json |
SSM param holding the credentials JSON |
drive_folder (tool arg) |
Gmail Attachments |
Destination Drive folder for downloads |
Security
- The server can send mail as you. Every send is logged (recipients, subject, timestamp). There is no recipient allowlist by default; treat the tool as injection-sensitive.
- The refresh token never touches disk — only an SSM SecureString, fetched in-memory at launch.
- Scopes are least-privilege: read Gmail, send Gmail, and Drive access limited to app-created files.
Development
uv sync
uv run pytest # unit tests (all mocked; no live Google/AWS calls)
uv run ruff check . # lint
Layout: src/gmail_mcp/ — auth.py (SSM→credentials), gmail_client.py (send/search/attachments),
drive_client.py (folder + upload), attachments.py (orchestration), server.py (FastMCP wiring),
bootstrap.py (token mint).
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.