Fetch web pages, strip to clean readable text, summarize into agent-ready markdown. Research assistant foundation. No browser required.