What is the Zyte API rendering fallback?

When it is used

Every page is read first as its server sends it. Only a page that needs rendering is offered to the fallback:

• a page that answered normally but has hardly any visible words (60 unless SEOmise changes it) and shows the signs of a JavaScript app (an empty root or app element, Next.js or Nuxt data, Angular attributes) or is mostly script; or
• a bot wall (a challenge page answering "forbidden" or "unavailable") in front of your own site, and only when you have verified the domain (Domain verification). A request to slow down (HTTP 429) is never bypassed: the scan slows down instead.

Pages that read normally are never sent. A scan that already renders with SEOmise's own browser (Render JavaScript) does not use the fallback for that.

Your site's rules come first

Before any page is sent: your robots.txt must allow SEOmise to read it, the site must not be on our opt-out list, and the address must be a public web address. The fallback is never a way round a site's rules.

Limits

• The first pages of a scan that need rendering, 20 unless SEOmise changes it.
• A monthly number of pages per workspace (300 unless changed); after that, pages keep their plain read and the scan says so.
• Rendered pages are not stored as such: the scan reads its usual fields from them (title, description, headings, text, links) as it does for every page.

Where you see it

The scan's summary counts the pages read through the fallback next to the pages rendered by SEOmise's own browser. A page that could not be read keeps what the plain read found.