Provides a universal read path for agents to extract facts, metadata, and provenance from local files and guarded remote URLs without using generative LLMs.
MCP server enabling local-first web search, fetch, extract, and caching with citeable excerpts, no API key required. Supports research workflows for agents and apps.