Web-content reader: fetch any http(s) URL and return clean readable text, title/description/canonical, and outbound links as JSON — the data an agent ingests on each step of a crawl or RAG loop. Input: url.