A Model Context Protocol (MCP) server for Gmail integration in Claude Desktop with auto authentication support. This server enables AI assistants to manage Gmail through natural language interactions.
What is Gmail-MCP-Server
Gmail AutoAuth MCP Server
A Model Context Protocol (MCP) server for Gmail integration in Claude Desktop with auto authentication support. This server enables AI assistants to manage Gmail through natural language interactions.
Features
- Send emails with subject, content, attachments, and recipients
- Support for HTML emails and multipart messages with both HTML and plain text versions
- Full support for international characters in subject lines and email content
- Read email messages by ID with advanced MIME structure handling
- View email attachments information (filenames, types, sizes)
- Search emails with various criteria (subject, sender, date range)
- Comprehensive label management with ability to create, update, delete and list labels
- List all available Gmail labels (system and user-defined)
- List emails in inbox, sent, or custom labels
- Mark emails as read/unread
- Move emails to different labels/folders
- Delete emails
- Batch operations for efficiently processing multiple emails at once
- Full integration with Gmail API
- Simple OAuth2 authentication flow with auto browser launch
- Support for both Desktop and Web application credentials
- Global credential storage for convenience
Installation & Authentication
Installing via Smithery
To install Gmail AutoAuth for Claude Desktop automatically via Smithery:
npx -y @smithery/cli install @gongrzhe/server-gmail-autoauth-mcp --client claude
Installing Manually
-
Create a Google Cloud Project and obtain credentials:
a. Create a Google Cloud Project:
- Go to Google Cloud Console
- Create a new project or select an existing one
- Enable the Gmail API for your project
b. Create OAuth 2.0 Credentials:
- Go to "APIs & Services" > "Credentials"
- Click "Create Credentials" > "OAuth client ID"
- Choose either "Desktop app" or "Web application" as application type
- Give it a name and click "Create"
- For Web application, add
http://localhost:3000/oauth2callback
to the authorized redirect URIs - Download the JSON file of your client's OAuth keys
- Rename the key file to
gcp-oauth.keys.json
-
Run Authentication:
You can authenticate in two ways:
a. Global Authentication (Recommended):
# First time: Place gcp-oauth.keys.json in your home directory's .gmail-mcp folder mkdir -p ~/.gmail-mcp mv gcp-oauth.keys.json ~/.gmail-mcp/ # Run authentication from anywhere npx @gongrzhe/server-gmail-autoauth-mcp auth
b. Local Authentication:
# Place gcp-oauth.keys.json in your current directory # The file will be automatically copied to global config npx @gongrzhe/server-gmail-autoauth-mcp auth
The authentication process will:
- Look for
gcp-oauth.keys.json
in the current directory or~/.gmail-mcp/
- If found in current directory, copy it to
~/.gmail-mcp/
- Open your default browser for Google authentication
- Save credentials as
~/.gmail-mcp/credentials.json
Note:
- After successful authentication, credentials are stored globally in
~/.gmail-mcp/
and can be used from any directory - Both Desktop app and Web application credentials are supported
- For Web application credentials, make sure to add
http://localhost:3000/oauth2callback
to your authorized redirect URIs
- Look for
-
Configure in Claude Desktop:
{
"mcpServers": {
"gmail": {
"command": "npx",
"args": [
"@gongrzhe/server-gmail-autoauth-mcp"
]
}
}
}
Docker Support
If you prefer using Docker:
- Authentication:
docker run -i --rm \
--mount type=bind,source=/path/to/gcp-oauth.keys.json,target=/gcp-oauth.keys.json \
-v mcp-gmail:/gmail-server \
-e GMAIL_OAUTH_PATH=/gcp-oauth.keys.json \
-e "GMAIL_CREDENTIALS_PATH=/gmail-server/credentials.json" \
-p 3000:3000 \
mcp/gmail auth
- Usage:
{
"mcpServers": {
"gmail": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-v",
"mcp-gmail:/gmail-server",
"-e",
"GMAIL_CREDENTIALS_PATH=/gmail-server/credentials.json",
"mcp/gmail"
]
}
}
}
Cloud Server Authentication
For cloud server environments (like n8n), you can specify a custom callback URL during authentication:
npx @gongrzhe/server-gmail-autoauth-mcp auth https://gmail.gongrzhe.com/oauth2callback
Setup Instructions for Cloud Environment
-
Configure Reverse Proxy:
- Set up your n8n container to expose a port for authentication
- Configure a reverse proxy to forward traffic from your domain (e.g.,
gmail.gongrzhe.com
) to this port
-
DNS Configuration:
- Add an A record in your DNS settings to resolve your domain to your cloud server's IP address
-
Google Cloud Platform Setup:
- In your Google Cloud Console, add your custom domain callback URL (e.g.,
https://gmail.gongrzhe.com/oauth2callback
) to the authorized redirect URIs list
- In your Google Cloud Console, add your custom domain callback URL (e.g.,
-
Run Authentication:
npx @gongrzhe/server-gmail-autoauth-mcp auth https://gmail.gongrzhe.com/oauth2callback
-
Configure in your application:
{ "mcpServers": { "gmail": { "command": "npx", "args": [ "@gongrzhe/server-gmail-autoauth-mcp" ] } } }
This approach allows authentication flows to work properly in environments where localhost isn't accessible, such as containerized applications or cloud servers.
Available Tools
The server provides the following tools that can be used through Claude Desktop:
1. Send Email (send_email
)
Sends a new email immediately. Supports plain text, HTML, or multipart emails.
{
"to": ["[email protected]"],
"subject": "Meeting Tomorrow",
"body": "Hi,\n\nJust a reminder about our meeting tomorrow at 10 AM.\n\nBest regards",
"cc": ["[email protected]"],
"bcc": ["[email protected]"],
"mimeType": "text/plain"
}
HTML Email Example:
{
"to": ["[email protected]"],
"subject": "Meeting Tomorrow",
"mimeType": "text/html",
"body": "<html><body><h1>Meeting Reminder</h1><p>Just a reminder about our <b>meeting tomorrow</b> at 10 AM.</p><p>Best regards</p></body></html>"
}
Multipart Email Example (HTML + Plain Text):
{
"to": ["[email protected]"],
"subject": "Meeting Tomorrow",
"mimeType": "multipart/alternative",
"body": "Hi,\n\nJust a reminder about our meeting tomorrow at 10 AM.\n\nBest regards",
"htmlBody": "<html><body><h1>Meeting Reminder</h1><p>Just a reminder about our <b>meeting tomorrow</b> at 10 AM.</p><p>Best regards</p></body></html>"
}
2. Draft Email (draft_email
)
Creates a draft email without sending it.
{
"to": ["[email protected]"],
"subject": "Draft Report",
"body": "Here's the draft report for your review.",
"cc": ["[email protected]"]
}
3. Read Email (read_email
)
Retrieves the content of a specific email by its ID.
{
"messageId": "182ab45cd67ef"
}
4. Search Emails (search_emails
)
Searches for emails using Gmail search syntax.
{
"query": "from:[email protected] after:2024/01/01 has:attachment",
"maxResults": 10
}
5. Modify Email (modify_email
)
Adds or removes labels from emails (move to different folders, archive, etc.).
{
"messageId": "182ab45cd67ef",
"addLabelIds": ["IMPORTANT"],
"removeLabelIds": ["INBOX"]
}
6. Delete Email (delete_email
)
Permanently deletes an email.
{
"messageId": "182ab45cd67ef"
}
7. List Email Labels (list_email_labels
)
Retrieves all available Gmail labels.
{}
8. Create Label (create_label
)
Creates a new Gmail label.
{
"name": "Important Projects",
"messageListVisibility": "show",
"labelListVisibility": "labelShow"
}
9. Update Label (update_label
)
Updates an existing Gmail label.
{
"id": "Label_1234567890",
"name": "Urgent Projects",
"messageListVisibility": "show",
"labelListVisibility": "labelShow"
}
10. Delete Label (delete_label
)
Deletes a Gmail label.
{
"id": "Label_1234567890"
}
11. Get or Create Label (get_or_create_label
)
Gets an existing label by name or creates it if it doesn't exist.
{
"name": "Project XYZ",
"messageListVisibility": "show",
"labelListVisibility": "labelShow"
}
12. Batch Modify Emails (batch_modify_emails
)
Modifies labels for multiple emails in efficient batches.
{
"messageIds": ["182ab45cd67ef", "182ab45cd67eg", "182ab45cd67eh"],
"addLabelIds": ["IMPORTANT"],
"removeLabelIds": ["INBOX"],
"batchSize": 50
}
13. Batch Delete Emails (batch_delete_emails
)
Permanently deletes multiple emails in efficient batches.
{
"messageIds": ["182ab45cd67ef", "182ab45cd67eg", "182ab45cd67eh"],
"batchSize": 50
}
Advanced Search Syntax
The search_emails
tool supports Gmail's powerful search operators:
Operator | Example | Description |
---|---|---|
from: |
from:[email protected] |
Emails from a specific sender |
to: |
to:[email protected] |
Emails sent to a specific recipient |
subject: |
subject:"meeting notes" |
Emails with specific text in the subject |
has:attachment |
has:attachment |
Emails with attachments |
after: |
after:2024/01/01 |
Emails received after a date |
before: |
before:2024/02/01 |
Emails received before a date |
is: |
is:unread |
Emails with a specific state |
label: |
label:work |
Emails with a specific label |
You can combine multiple operators: from:[email protected] after:2024/01/01 has:attachment
Advanced Features
Email Content Extraction
The server intelligently extracts email content from complex MIME structures:
- Prioritizes plain text content when available
- Falls back to HTML content if plain text is not available
- Handles multi-part MIME messages with nested parts
- Processes attachments information (filename, type, size)
- Preserves original email headers (From, To, Subject, Date)
International Character Support
The server fully supports non-ASCII characters in email subjects and content, including:
- Turkish, Chinese, Japanese, Korean, and other non-Latin alphabets
- Special characters and symbols
- Proper encoding ensures correct display in email clients
Comprehensive Label Management
The server provides a complete set of tools for managing Gmail labels:
- Create Labels: Create new labels with customizable visibility settings
- Update Labels: Rename labels or change their visibility settings
- Delete Labels: Remove user-created labels (system labels are protected)
- Find or Create: Get a label by name or automatically create it if not found
- List All Labels: View all system and user labels with detailed information
- Label Visibility Options: Control how labels appear in message and label lists
Label visibility settings include:
messageListVisibility
: Controls whether the label appears in the message list (show
orhide
)labelListVisibility
: Controls how the label appears in the label list (labelShow
,labelShowIfUnread
, orlabelHide
)
These label management features enable sophisticated organization of emails directly through Claude, without needing to switch to the Gmail interface.
Batch Operations
The server includes efficient batch processing capabilities:
- Process up to 50 emails at once (configurable batch size)
- Automatic chunking of large email sets to avoid API limits
- Detailed success/failure reporting for each operation
- Graceful error handling with individual retries
- Perfect for bulk inbox management and organization tasks
Security Notes
- OAuth credentials are stored securely in your local environment (
~/.gmail-mcp/
) - The server uses offline access to maintain persistent authentication
- Never share or commit your credentials to version control
- Regularly review and revoke unused access in your Google Account settings
- Credentials are stored globally but are only accessible by the current user
Troubleshooting
-
OAuth Keys Not Found
- Make sure
gcp-oauth.keys.json
is in either your current directory or~/.gmail-mcp/
- Check file permissions
- Make sure
-
Invalid Credentials Format
- Ensure your OAuth keys file contains either
web
orinstalled
credentials - For web applications, verify the redirect URI is correctly configured
- Ensure your OAuth keys file contains either
-
Port Already in Use
- If port 3000 is already in use, please free it up before running authentication
- You can find and stop the process using that port
-
Batch Operation Failures
- If batch operations fail, they automatically retry individual items
- Check the detailed error messages for specific failures
- Consider reducing the batch size if you encounter rate limiting
Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
Running evals
The evals package loads an mcp client that then runs the index.ts file, so there is no need to rebuild between tests. You can load environment variables by prefixing the npx command. Full documentation can be found here.
OPENAI_API_KEY=your-key npx mcp-eval src/evals/evals.ts src/index.ts
License
MIT
Support
If you encounter any issues or have questions, please file an issue on the GitHub repository.
Leave a Comment
Frequently Asked Questions
What is MCP?
MCP (Model Context Protocol) is an open protocol that standardizes how applications provide context to LLMs. Think of MCP like a USB-C port for AI applications, providing a standardized way to connect AI models to different data sources and tools.
What are MCP Servers?
MCP Servers are lightweight programs that expose specific capabilities through the standardized Model Context Protocol. They act as bridges between LLMs like Claude and various data sources or services, allowing secure access to files, databases, APIs, and other resources.
How do MCP Servers work?
MCP Servers follow a client-server architecture where a host application (like Claude Desktop) connects to multiple servers. Each server provides specific functionality through standardized endpoints and protocols, enabling Claude to access data and perform actions through the standardized protocol.
Are MCP Servers secure?
Yes, MCP Servers are designed with security in mind. They run locally with explicit configuration and permissions, require user approval for actions, and include built-in security features to prevent unauthorized access and ensure data privacy.
Related MCP Servers
sylphlab pdf reader mcp
An MCP server built with Node.js/TypeScript that allows AI agents to securely read PDF files (local or URL) and extract text, metadata, or page counts. Uses pdf-parse.
aashari mcp server atlassian bitbucket
Node.js/TypeScript MCP server for Atlassian Bitbucket. Enables AI systems (LLMs) to interact with workspaces, repositories, and pull requests via tools (list, get, comment, search). Connects AI directly to version control workflows through the standard MCP interface.
aashari mcp server atlassian confluence
Node.js/TypeScript MCP server for Atlassian Confluence. Provides tools enabling AI systems (LLMs) to list/get spaces & pages (content formatted as Markdown) and search via CQL. Connects AI seamlessly to Confluence knowledge bases using the standard MCP interface.
weibaohui k8m
一款轻量级、跨平台的 Mini Kubernetes AI Dashboard,支持大模型+智能体+MCP(支持设置操作权限),集成多集群管理、智能分析、实时异常检测等功能,支持多架构并可单文件部署,助力高效集群管理与运维优化。
watchdealer pavel watchbase mcp server
MCP Server for structured and standardized querying of watch-related metadata such as brands, families, and reference details from WatchBase.com.
vgnshiyer apple books mcp
Apple Books MCP Server
ttiimmaacc cinema4d mcp
Cinema 4D plugin integrating Claude AI for prompt-driven 3D modeling, scene creation, and manipulation.
sv mcp paradex py
Connect AI agents to the Paradex trading platform. Retrieve market data, manage accounts, and execute trades seamlessly. Enhance your trading experience with automated tools and real-time insights.
SamllPigYanDong revit mcp
Revit MCP. A Model Context Protocol server for Revit integration, enabling seamless communication between Claude AI and Autodesk Revit.
Rayyan9477 linkedin mcp
A powerful Model Context Protocol server for LinkedIn interactions that enables AI assistants to search for jobs, generate resumes and cover letters, and manage job applications programmatically.
Submit Your MCP Server
Share your MCP server with the community
Submit Now