Extracts structured content from any public URL: page title, meta description, H1-H3 headings, all links (with text and internal/external flag), and a 500-character text preview. Useful for research agents following link chains from on-chain data, auditing page structure, or seeding downstream text-generation calls.