• responseOptions.onlyMainContent: true serves only the page’s main content, without navigation, headers, footers and sidebars, in either format. GET /{url} always serves the whole page.
  • The content is found with the Readability.js algorithm (a Go port), on a best-effort basis: an article extracts well, a list or index page poorly. Where it falls short, run your own extractor on the plain html.
  • The page’s title, when one is found, leads the content as an <h1> in html and a # heading in markdown. Links and images are made absolute against the final URL.
  • A page with no readable content serves an empty html or markdown, not an error.
  • The 10 MiB cap applies to the content as served, and to the html it is extracted from.