Search From a Page URL

Fetches the page at the supplied URL, identifies the single company, product, or person the page is primarily about, and returns the federated entity, person, and news results for that name together with the synthesized query itself. A page whose content cannot be read returns 404; a fetch or inference failure returns a problem (503 only when a retry can succeed), never an empty result. For direct URL-to-record identity resolution use the entities lookup operation with the url instead. Send Prefer: respond-async to skip the live wait: the search is committed as a link search job and the read answers 202 with a Location status URL and Retry-After; poll it. A repeat of a search that completed in the last few minutes answers 200 from its stored result. A failed job commit answers 503; a caller already holding the maximum active link searches gets 429.