Nvillaluenga/mcp-server-with-google-auth

MCP Server Google Auth Project

This project demonstrates an MCP (Model Context Protocol) server with Google Drive integration and a client to interact with it using Gemini.

Prerequisites

• Python 3.12+
• uv (Python package installer and virtual environment manager)
• Google Cloud Project with Vertex AI API enabled.
• OAuth 2.0 credentials.json file from Google Cloud Console.

Setup

1. Clone the repository:

   git clone https://github.com/Nvillaluenga/mcp-server-with-google-auth.git
   cd mcp-server-with-google-auth
   

2. Create and activate a virtual environment with uv:

   uv venv
   source .venv/bin/activate  # On Windows: .venv\Scripts\activate
   

3. Install dependencies with uv sync from pyproject.toml:

   uv sync
   

   The dependencies are defined in pyproject.toml and include:

   • fastapi - Web framework for the MCP server
   • fastmcp - MCP server implementation
   • google-api-python-client - Google API client
   • google-auth and google-auth-oauthlib - Google authentication
   • google-genai - Google Generative AI interface

4. Set up Google OAuth Credentials:

   • Go to the Google Cloud Console
   • Navigate to "APIs & Services" > "Credentials"
   • Create OAuth 2.0 Client IDs (Web application)
   • Add http://localhost:8081 to "Authorized JavaScript origins"
   • Add http://localhost:8081/oauth2callback to "Authorized redirect URIs"
   • Download and rename the credentials file to credentials.json and place it in the project root

5. Configure Environment Variables:

   • Create .env from the template:

     cp .env.template .env
     

   • Edit .env with your configuration:
     • PROJECT: Your Google Cloud Project ID
     • LOCATION: Your GCP region (e.g., us-central1)
     • DEFAULT_MODEL: Gemini model name (e.g., gemini-2.0-flash-001)
     • HOST: Server host (default: localhost)
     • PORT: Server port (default: 8081)

Running the Application

You'll need two terminals with active uv environments:

1. Start the MCP Server:

python drive_web_mcp_server.py

2. Run the Client:

python main.py

Or specify a server URL:

python main.py http://your-server-address:port

The client will prompt you to authenticate with Google Drive:

• Open the displayed URL in your browser
• Complete the Google authentication flow
• Return to the terminal and start querying Gemini with access to Google Drive

How it Works

• drive_web_mcp_server.py: FastAPI-based MCP server that:

  • Handles OAuth 2.0 authentication with Google
  • Exposes tools for the LLM to call
  • Stores client credentials in memory by client_id

• mcp_client.py: Contains MCPClient and GeminiMCPClient classes:

  • Manages connections to MCP servers
  • Handles authentication flow
  • Interfaces with Gemini models
  • Translates between MCP tools and Gemini function calls

• main.py: CLI to interact with GeminiMCPClient

• .env and .env.template: Environment configuration

• credentials.json: Google OAuth 2.0 client secrets (never commit this)

Authentication Flow Diagram

The following diagram illustrates the authentication process between the MCP client, MCP server and Google OAuth:

Available Tools

• search_drive_files(query: str): Search files in Google Drive
• check_authentication_status(): Check client authentication status