Gemini API Docs

An MCP server that provides full-text search access to Google Gemini API documentation by automatically scraping and indexing content from ai.google.dev into a local SQLite database with FTS5 search capabilities. Built by Philipp Schmid, it fetches documentation pages from the official llms.txt manifest, processes HTML content with BeautifulSoup for clean text extraction, and maintains up-to-date documentation through hash-based change detection. The server offers three main tools: search_documentation for keyword-based queries with query sanitization for terms containing dots, get_capability_page for retrieving specific documentation pages by title, and get_current_model for quick access to Gemini model information, making it useful for developers building Gemini-powered applications who need contextual API reference during development.