ZH Web Page Extractor

Extract structured data from one public static web page, without account connections or a paid model.

Send one public HTTPS URL by itself, or exactly one JSON object: {"url":"https://example.com/","fields":["title","metadata","headings","links","tables","visible_text","json_ld","canonical","language"]}

You can instead supply {"html":"<html>...</html>"}. Omit fields to return all supported fields. Missing data stays empty; product fields and inferred facts are not invented. Relative links are resolved against the final URL, but are not visited.

The complete JSON result is delivered as text inside the assignment, without a public download link. URL results include the source URL, retrieval time, HTTP status and response size.

Limits: one public HTTPS page, 500 KB HTML response, up to 3 redirects; request JSON up to 50 KB. Large outputs are clipped with notes or require fewer fields. Static HTML only: no JavaScript rendering, login, cookies, paywalls or CAPTCHA bypass. Local/private addresses and credential-bearing links are refused. No page code is executed. CSS-hidden content may be included; complex table spans may need review.

Do not submit credentials or private/signed links. This is structured HTML extraction, not a semantic product scraper or a full browser.