Gemini MCP Server

An MCP (Model Context Protocol) server implementation that enables using Google’s Gemini AI models through Claude or other MCP clients.

Features

• Full MCP protocol support via SSE transport
• Gemini 1.5 Pro model integration
• Sequential thinking tool implementation
• Environment variable configuration

Deployment on Smithery AI

This server is designed to be easily deployed on Smithery AI. Once deployed, you’ll be able to connect to it from any MCP client like Claude.

Configuration

The server requires the following environment variables:

• GEMINI_API_KEY: Your Google AI Studio API key for accessing Gemini models

Local Development

If you want to run this server locally:

1. Clone the repository
2. Install dependencies: npm install
3. Set your Gemini API key: export GEMINI_API_KEY=your_key_here
4. Start the server: npm start

How to Use

After deploying to Smithery AI, you’ll get a unique URL that you can add to your MCP client:

1. Copy the deployment URL from Smithery AI
2. In Claude Desktop, go to Settings > MCP Servers
3. Add a new server with:
   • Name: Gemini MCP
   • URL: Your Smithery deployment URL
4. Save and restart Claude

License

MIT