Gemini Docs MCP Server

# Gemini Docs MCP Server This is an MCP (Model Context Protocol) server that provides tools to search and fetch Google's Gemini API documentation at https://ai.google.dev/gemini-api/docs. ## Project Structure ``` gemini-docs-mcp/ ├── package.json # Node.js dependencies and scripts ├── tsconfig.json # TypeScript configuration ├── README.md # User documentation ├── CLAUDE.md # This file ├── src/ │ ├── index.ts # MCP server entry point with tool registrations │ ├── constants.ts # Configuration constants and documentation index │ ├── test-mcp.ts # E2E tests for the tools │ └── utils/ │ ├── parser.ts # HTML fetching and parsing utilities │ └── search.ts # Search functionality └── dist/ # Compiled JavaScript (npm run build) ``` ## Tools Provided 1. **search_gemini_docs** - Searches the documentation index for relevant pages 2. **fetch_gemini_doc** - Fetches and parses a specific documentation page ## Key Files - `src/index.ts` - Main MCP server with tool registrations using Zod schemas - `src/constants.ts` - Contains DOC_INDEX with all known documentation pages and their keywords - `src/utils/parser.ts` - Uses undici for HTTP and cheerio for HTML parsing - `src/utils/search.ts` - Keyword-based search against the documentation index ## Development Commands ```bash npm install # Install dependencies npm run build # Compile TypeScript npm run dev # Development mode with hot reload npm run test # Run E2E tests npm start # Run the built server ``` ## Testing Run `npm run test` to execute the E2E tests in `src/test-mcp.ts`. All 8 tests should pass. ## Notes - The search uses a static index defined in `constants.ts` (not live search) - Fetching pages makes real HTTP requests to ai.google.dev - The server uses stdio transport for MCP communication