What it checks

Every check rests on an engine's own documentation or a published measurement. Things with no such basis, such as structured data or llms.txt, are not checked.

Blocked for AI

The engine's documentation says the page can't appear in its answers.

ChatGPT is blocked (robots.txt)

robots.txt disallows the page for OAI-SearchBot. OpenAI says ChatGPT search only shows pages its search bot may read (OpenAI crawlers).

Fix: Allow OAI-SearchBot on these paths in robots.txt.

Google is blocked (robots.txt)

robots.txt disallows the page for Googlebot. Google says AI Overviews only show pages Googlebot may read (AI features and your website, robots.txt).

Fix: Allow Googlebot on these paths in robots.txt.

Tells Google to skip it (noindex)

A robots meta tag for robots or googlebot, or the X-Robots-Tag header, says noindex or none, or Search Console reports the page blocked by one. Google says a noindex page stays out of Search, so AI Overviews can't show it (Block indexing).

Fix: Remove noindex from the robots meta tag or the X-Robots-Tag header. On Shopify it usually comes from an SEO app or theme.liquid.

Hides its text from Google (nosnippet)

The same tags say nosnippet or max-snippet:0. Google says AI Overviews only quote pages that let Google show their text (Robots meta tags).

Fix: Remove nosnippet and max-snippet:0 from the robots meta tag or the X-Robots-Tag header.

Page is missing (404)

The page answers 404 or 410. A missing page drops out of Google, and answers that cite it send people to an error (Technical requirements).

Fix: Bring the page back, or redirect the address to the page that replaced it.

Not on Google Search

Needs Search Console. Google reports the page as not indexed or as a soft 404, and robots.txt, noindex or the page's own canonical don't already explain it. Google says AI Overviews only use pages in its index (AI features and your website).

Fix: In Search Console, paste the page's address into the search bar at the top to see why Google left it out (URL Inspection). Fix that, then click Request indexing.

May hold AI back

The documentation names an effect on crawling or indexing, or a measurement shows one.

robots.txt doesn't load

robots.txt answers a 5xx error or doesn't answer. Google says that while robots.txt fails, it stops reading your whole site (How Google reads robots.txt, RFC 9309).

Fix: Make robots.txt answer 200, or 404 if you have none.

Points Google to another page (canonical)

With Search Console, Google chose a different canonical than yours. Without it, the page's canonical points to another address. Google shows the page this one points to instead, so this one rarely appears (Canonical URLs).

Fix: Point the canonical at the page itself, unless the other page is the one to show.

No sitemap found

No sitemap in robots.txt, in Search Console or at the usual addresses such as /sitemap.xml. Google says a sitemap helps it find all of your pages (Sitemaps).

Fix: Publish a sitemap and name it in robots.txt with a Sitemap: line.

A sitemap doesn't load

A sitemap doesn't answer 200. Search engines can't use a sitemap that doesn't load (Sitemaps).

Fix: Make the sitemap answer 200, or remove it from robots.txt and Search Console.

Redirects to your homepage

An address other than the homepage redirects to the homepage. Google treats an old address that lands on your homepage like a missing page (Site moves).

Fix: Redirect each old address to the page closest to it, not the homepage.

AI answers show another country's page

A page AI answers used is made for another market: its path starts with another locale, such as /en-gb/. Shoppers in this market land on a page meant for another country (Localized versions).

Fix: Link the market versions with hreflang, so Google can show this market its own page.

Almost no text AI can read

The HTML the server sends has fewer than 50 words of the page's own text: the <main> element, or else the text that isn't on your homepage. Vercel measured that ChatGPT's and Claude's crawlers don't run JavaScript, so they see almost nothing here (The rise of the AI crawler).

Fix: Put the page's main text in the HTML the server sends, not only in JavaScript.

Couldn't check

Your server answered SearchD with an error or a bot check, so the audit can't tell what AI bots get and doesn't judge the page.

Pages we couldn't open

The page answered 401, 403 or 429, a 5xx error after retries, or a bot check; it timed out or didn't connect; or your robots.txt disallows SearchD. If AI bots get the same answer, these pages can't show up in AI answers.

Fix: Let SearchD through, then run the audit again. SearchD reads with the user agent SearchD/1.0 (+https://searchd.ai).