MCP Gemini Server
@bsmi021
This project provides a dedicated MCP (Model Context Protocol) server that wraps the @google/genai SDK. It exposes Google's Gemini model capabilities as standard MCP tools, allowing other LLMs (like Cline) or MCP-compatible systems to leverage Gemini's features as a backend workh
Overview
What is MCP Gemini Server?
MCP Gemini Server wraps the @google/genai SDK to expose Google Gemini models as standard MCP (Model Context Protocol) tools. It allows other LLMs or MCP-compatible systems to leverage Gemini’s generation, function calling, chat, caching, and URL-based multimedia analysis capabilities.
How to use MCP Gemini Server?
Install Node.js v18+, clone the repository, run npm install and npm run build, then generate a connection token. Configure your MCP client’s settings (e.g., cline_mcp_settings.json) with the absolute path to dist/server.js, required environment variables (GOOGLE_GEMINI_API_KEY, MCP_SERVER_HOST, MCP_SERVER_PORT, MCP_CONNECTION_TOKEN), and optional ones like GOOGLE_GEMINI_MODEL. Restart your MCP client to begin using the tools.
Key features of MCP Gemini Server
- Text generation and streaming with system instructions
- Stateful chat with context across multiple turns
- URL‑based image and YouTube video analysis
- Caching support for optimized prompts
- Image generation (Gemini 2.0 Flash and Imagen 3.1)
- MCP client to connect to other MCP servers
Use cases of MCP Gemini Server
- Analyze images hosted at public URLs without file uploads
- Summarize or extract insights from public YouTube videos
- Build conversational AI with persistent chat history
- Generate images from text prompts using Gemini models
- Orchestrate multiple MCP servers via the built‑in MCP client
FAQ from MCP Gemini Server
Does it support direct file uploads or base64 images?
No. The server only processes multimedia from publicly accessible URLs (images, YouTube videos, web content). Local files, base64 data, and audio uploads are not supported.
What Gemini models are available?
It supports gemini-1.5-pro-latest, gemini-1.5-flash, and gemini-2.5-pro. Image generation requires Gemini 2.0 Flash Experimental or Imagen 3.1.
Do I need a Google AI Studio API key?
Yes, a valid API key from Google AI Studio is required. Vertex AI authentication is not currently supported.
What are the runtime and transport requirements?
Requires Node.js v18 or later. The server uses a standard HTTP transport and must be configured with MCP_SERVER_HOST, MCP_SERVER_PORT, and a MCP_CONNECTION_TOKEN for secure communication.
Can I cache prompts to reduce costs?
Yes, the server provides tools to create, list, retrieve, update, and delete cached content. The Caching API works only with Google AI Studio API keys.
