free-search-mcp
View on GitHubLocal-first, no-API-key MCP server for web search, fetch, and document reading. Multi-engine (DuckDuckGo, Mojeek, Startpage), smart Playwright fallback, FTS5 cache, LLM-tuned output.
A Python MCP server that gives AI clients keyless, multi-engine web search, page fetching, and document reading. Includes a research tool, browser-rendered fallback, structured extraction, and an FTS5 cache.
Use Cases
Search the web from an AI assistantFetch and extract readable page contentResearch a question with a combined search-and-fetch callRead PDFs, office files, and other documentsCompare multiple web pagesSearch previously fetched pagesExplore academic papers and citation graphs
Built With
- Language
- Python
- Frameworks
- Model Context Protocol · Playwright · Trafilatura
Tags
web search · web fetching · document reading · multi-engine search · local-first · no API key · browser automation · FTS5 cache · research · MCP server