Google Researcher MCP Server

What is Google Researcher MCP Server?

Google Researcher MCP Server gives AI assistants real-time web research abilities via Google Search, content scraping, and Gemini AI analysis. It implements the Model Context Protocol (MCP) and uses persistent caching to reduce API costs and improve performance. The server is built for developers integrating research capabilities into MCP‑compatible AI clients.

How to use Google Researcher MCP Server?

Clone the repository, install dependencies, copy .env.example to .env, and fill in your Google Custom Search API key, Custom Search Engine ID, and Gemini API key. Run in development mode with npm run dev or build and run production with npm run build && npm start. The server provides four tools – google_search, scrape_page, analyze_with_gemini, and research_topic – callable over STDIO or HTTP+SSE transports.

Key features of Google Researcher MCP Server

• Four research tools: search, scrape, analyze, and combined workflow
• Persistent caching (memory + disk) for repeated queries
• Session resumption support for web clients
• Dual transport: STDIO and HTTP+SSE
• Management API endpoints for cache and event store statistics
• OAuth 2.1 authorization with granular scope control

Use cases of Google Researcher MCP Server

• Enable an AI coding assistant to look up current web information during development
• Automate research workflows that collect, scrape, and summarize multiple sources
• Build a conversational agent that answers questions using real‑time search results
• Reduce API costs by caching identical search and analysis requests

FAQ from Google Researcher MCP Server

What runtime and API keys does the server require?

Node.js v18 or later, plus API keys for Google Custom Search, a Google Custom Search Engine ID, and a Google Gemini API key.

How do I run the server for development?

Use npm run dev for automatic reloading on file changes, or build and start with npm run build && npm start for production.

What transport options are available?

The server supports STDIO (direct process) and HTTP+SSE (web) transports. HTTP+SSE requires OAuth 2.1 authentication with a Bearer token.

Can I monitor or clear the cache?

Yes. HTTP management endpoints include GET /mcp/cache-stats, POST /mcp/cache-invalidate, and POST /mcp/cache-persist, each protected by appropriate OAuth scopes.

How do I integrate this server with Roo Code?

Create a .roo/mcp.json file in your project with the server’s path and environment variables, using STDIO transport. The README provides a complete example configuration.