Google Drive MCP (Model Context Protocol)

This project consists of two separate NestJS servers:

1. Auth Server - Handles Google OAuth2 authentication
2. Drive Server - Handles Google Drive CRUD operations

Prerequisites

• Node.js (v14 or higher)
• npm or yarn
• Google Cloud Platform account with Google Drive API enabled
• OAuth 2.0 credentials from Google Cloud Console

Setup

1. Clone the repository
2. Install dependencies:

   npm install

1. Create a .env file in the root directory with the following variables:

   # OAuth Server Configuration
   AUTH_SERVER_PORT=3000
   AUTH_SERVER_URL=http://localhost:3000

   # Drive Server Configuration
   DRIVE_SERVER_PORT=3001
   DRIVE_SERVER_URL=http://localhost:3001

   # Google OAuth Configuration
   GOOGLE_CLIENT_ID=your_client_id_here
   GOOGLE_CLIENT_SECRET=your_client_secret_here
   GOOGLE_REDIRECT_URI=http://localhost:3000/auth/google/callback

   # Session Configuration
   SESSION_SECRET=your_session_secret_here

Running the Servers

1. Start the Auth Server:

   npm run start:auth

1. Start the Drive Server:

   npm run start:drive

API Endpoints

Auth Server (Port 3000)

• GET /auth/google - Get Google OAuth URL
• GET /auth/google/callback - OAuth callback endpoint
• GET /auth/tokens - Get current user's tokens
• POST /auth/logout - Logout endpoint

Drive Server (Port 3001)

• GET /drive/files - List files
• Query params: pageSize, pageToken
• GET /drive/files/:fileId - Get file details
• POST /drive/files - Create new file
• Body: file (multipart/form-data), metadata (JSON)
• PUT /drive/files/:fileId - Update file
• Body: file (multipart/form-data), metadata (JSON)
• DELETE /drive/files/:fileId - Delete file

Development

For development with hot-reload:

# Auth Server
npm run dev:auth

# Drive Server
npm run dev:drive

Security Considerations

1. Always use HTTPS in production
2. Implement proper session management
3. Store tokens securely
4. Implement rate limiting
5. Add proper error handling
6. Add request validation
7. Implement proper logging

Using with Claude

This MCP server allows Claude to access and interact with your Google Drive files. When properly configured, Claude can search, list, and read your files directly.

First, build the MCP server:

npm run build:mcp

Configuration for Claude Desktop

Add the following configuration to your Claude Desktop configuration file:

{
  "mcpServers": {
    "google-drive-mcp": {
      "command": "node",
      "args": ["/path/to/google-drive-mcp/dist/index.js"],
      "env": {
        "GDRIVE_CREDS_DIR": "/path/to/config/directory",
        "CLIENT_ID": "your-google-client-id",
        "CLIENT_SECRET": "your-google-client-secret"
      },
      "description": "This server connects Claude to Google Drive, allowing it to search, list, and read your files. You can ask Claude to find specific documents, read file contents, or work with data from your Google Drive."
    }
  }
}

Integration with Claude Code

This project is optimized for use with Claude Code, Anthropic's CLI tool. The combination enables powerful Google Drive automation and organization workflows.

Key Features for Claude Code

• Bulk file operations: Delete empty folders, organize files, batch processing
• Intelligent categorization: Automated folder sorting using naming patterns
• Parallel processing: Efficient handling of large datasets with background jobs
• Token management: Optimized API usage to conserve Claude's context window

Example Workflows

1. Mass File Organization:

   # Claude Code can help organize hundreds of files into categories
   # Using parallel API calls and intelligent categorization

1. Empty Folder Cleanup:

   # Automatically identify and remove empty folders
   # With verification and batch processing

1. Folder Structure Analysis:

   # Analyze folder hierarchies and recommend optimizations
   # Generate reports on file distribution and usage patterns

Best Practices

The CLAUDE.md file contains detailed best practices for:

• Token conservation strategies
• Memory management techniques
• Parallel processing patterns
• API efficiency optimization

These patterns were developed through real-world usage organizing 300+ folders and are designed to minimize context usage while maximizing automation capabilities.