Whether AI assistants (ChatGPT, Claude, Perplexity, Gemini, Copilot) can read and quote your site: AI crawler access, llms.txt and answer-ready content.
SmooxSEO runs 18 ai search checks on every page of your site:
A page can't be found, indexed or trusted, or search engines can't read part of the site. Fix it soon: it costs traffic.
robots.txt stops the robots that ChatGPT, Claude, Perplexity and others use to find answers. Your pages can't be quoted or linked in those answers.
Bing's robot is blocked, so the site can't appear in Bing, nor in Microsoft Copilot and the other assistants that answer with Bing results.
The page still ranks, but less well, or a real opportunity is lost. Fix it when you can.
Asked with their real names, these robots got an error or a challenge page while a normal browser got the site. Often Cloudflare's "Block AI bots" setting or a bot-protection rule. Tested from Smoox's server: real robots come from their own addresses, so treat it as a strong hint.
The heading asks a question, but the text under it doesn't answer it in its first sentence. AI answers quote short, direct answers, and readers want them too.
A rule for specific robots (<meta name="OAI-SearchBot" content="noindex"> or an X-Robots-Tag header naming them) keeps this page out of those assistants' answers.
Assistants that read pages live give up after a few seconds. This page took more than 5 seconds before the server started answering.
Good practice with a small effect, or something that is often intentional. Nice to have.
When someone pastes your link in ChatGPT, Claude or Perplexity, these robots fetch it. Blocked, the assistant can't read your page for that person.
GPTBot, ClaudeBot, Google-Extended and others can't use your text to train models. That's a legitimate choice: it doesn't stop you from being quoted in AI answers.
The home page never says plainly what the brand is ("Acme is a … that …"). Assistants describe you from sentences like this, and guess when there isn't one.
The page compares things in sentences. Tables are much easier for assistants (and readers) to extract.
No independent content was found outside the consent UI and navigation in the raw HTML. This does not prove a cookie wall or consent dependency.
Pages listed in llms.txt don't load, so an assistant following it hits dead ends.
No page answers your customers' common questions in one place. FAQs match how people ask assistants.
Specific facts (numbers, prices, dates, percentages) are quoted far more often than general statements.
People ask assistants questions. Headings that match those questions are easier to pick as an answer.
llms.txt is an optional map of the site for AI assistants, in plain text. No major assistant has said it uses it yet, so it's a small, cheap extra.
Phrases like "as mentioned above" break when an assistant quotes the paragraph on its own.
nosnippet, a small max-snippet or data-nosnippet tells Google not to quote the page: no text in search results and no use in AI Overviews.