Gemini MCP Server
Access Google's Gemini AI through the gemini-cli command-line tool.
Gemini MCP Server
A Model Context Protocol (MCP) server that provides AI-powered internet search capabilities through the gemini-cli tool.
Overview
This MCP server allows Claude and other AI assistants to perform real-time internet searches by interfacing with the gemini-cli command-line tool. It provides a single query tool that can search for current information and provide AI-powered analysis.
Features
- Real-time Search: Query current internet data through AI-powered search
- MCP Protocol: Implements the Model Context Protocol for seamless integration
- HTTP Streaming: Uses HTTP streaming transport for efficient communication
- TypeScript: Fully typed with TypeScript for reliability
- Graceful Shutdown: Handles SIGINT/SIGTERM signals for clean exits
Prerequisites
- Node.js 18+
gemini-cliinstalled and configured on your system- TypeScript (for development)
Installation
- Clone the repository:
git clone <repository-url>
cd gemini-mcp
- Install dependencies:
npm install
- Build the project:
npm run build
Usage
Development Mode
npm run dev
Production Mode
npm start
The server will start on port 13001 by default, or you can set a custom port:
PORT=8080 npm start
Configuration
Environment Variables
PORT: Server port (default: 13001)
MCP Tool
The server provides one tool:
query
- Description: Send a query to an AI agent that can search real-time internet data
- Parameters:
text(string): The search query or question to send to the AI agent
- Returns: AI-generated response with current information
Integration
With Claude Desktop
Add to your Claude Desktop configuration:
{
"mcpServers": {
"gemini-search": {
"command": "node",
"args": ["path/to/gemini-mcp/dist/index.js"],
"env": {
"PORT": "13001"
}
}
}
}
With Other MCP Clients
Connect to the HTTP streaming endpoint:
http://localhost:13001
Development
Scripts
npm run build- Compile TypeScript to JavaScriptnpm run start- Run the compiled servernpm run dev- Run in development mode with ts-node
Project Structure
src/
├── index.ts # Main server implementation
├── package.json # Project configuration
└── dist/ # Compiled JavaScript (after build)
Dependencies
- fastmcp: MCP server framework
- zod: Schema validation
- child_process: Node.js built-in for spawning processes
License
MIT
Contributing
- Fork the repository
- Create a feature branch
- Make your changes
- Test thoroughly
- Submit a pull request
Related Servers
Kintone OAuth MCP Server
A sample MCP server for kintone using OAuth, deployable on Cloudflare Workers.
MCP Salesforce Server
Provides seamless integration with Salesforce using OAuth authentication.
Netbird
List and analyze Netbird network peers, groups, policies, and more.
MCP Salesforce Connector
Interact with Salesforce data using SOQL queries and SOSL searches via an MCP server.
YouTube MCP Server
An MCP server for interacting with YouTube content, enabling AI models to access and manage YouTube data via its API.
Render MCP Server
Interact with Render resources via LLMs using the Render API.
RunPod MCP Server
Interact with the RunPod REST API to manage cloud GPU resources.
Remote MCP Server (Authless)
An authentication-free remote MCP server deployable on Cloudflare Workers or locally via npm.
DeepSeek
Access DeepSeek's advanced language models via the DeepSeek API.
Tradernet MCP Server
MCP server for Tradernet/Freedom24 trading API — portfolio, orders, quotes, alerts and more