AI Visibility Check
Can AI see and cite your website?
Paste your page's URL and find out whether ChatGPT, Claude, Perplexity and Google's AI Overviews can crawl, read and cite it - covering crawler access, structured data, how answerable your content is and whether it's citable. Every failing check's evidence is shown in full, free, in under a minute.
How it works
How does this tool work?
Paste a page's URL and this tool checks it the way an AI crawler would, then runs more than thirty checks that cover whether crawlers can reach the page, whether its structured data is machine-readable, whether it answers questions directly, whether it's citable, and whether the basic technical hygiene is in place. A score and every failing check's evidence come back in under a minute.
- Paste your homepage or any page's URL - no sign-up needed to see a score.
- The page is fetched the way an AI crawler would fetch it, then checked against crawler access, structured data, answerability and citation signals.
- You get a plain-language score and every failing check's evidence straight away, free.
- Priority-ordered fixes and paste-ready code unlock when you leave your email.
What it checks
Can AI crawlers reach your pages?
ChatGPT, Claude and Perplexity each run their own crawler, and Google splits ordinary indexing from AI-training access entirely, so one blanket robots.txt rule rarely covers all of them. A single misconfigured rule, a security plugin's default block, or a stray noindex tag can shut every one of them out at once, often without you ever noticing, because the page still looks completely normal to a human visitor browsing it directly.
This tool checks robots.txt against each named agent individually, includingOpenAI's GPTBot, OAI-SearchBot and ChatGPT-User, plus the equivalent Anthropic and Perplexity bots, because one blanket rule usually doesn't cover all of them.
What it checks
Is your content machine-readable?
AI systems don't read a page the way a person does. They lean on structured data, JSON-LD that marks up who runs your site, who wrote the page and what it is, to resolve ambiguity the surrounding prose leaves open. Missing, broken or dangling structured data means a crawler has to guess at facts this tool checks for directly instead.
This is the same reasoning that shapes howlarge language modelsweigh a source in the first place: an unambiguous, machine-readable version of a page's content and who wrote it counts for more than the same information left implicit in the prose.
What it checks
Does your page answer the question people ask?
Systems like ChatGPT and Perplexity extract answers rather than read essays, so a page that spends three paragraphs building up to its point is far harder to lift a clean answer from than one that states it plainly under a heading shaped like the question a reader typed in. This tool checks whether that direct-answer structure is there.
What it checks
Is your content something AI would want to cite?
Being cited depends on more than having the right information. AI systems weigh outbound links to credible sources, concrete numbers over vague claims, and a visible author byline as trust signals before treating a page as worth quoting. This tool checks for all three, because a page with no chosen sources looks the same to a crawler as one with nothing to say.
What it checks
Is the basic technical hygiene in place?
None of the above matters if the basics aren't in place first. A missing meta description, a canonical tag pointing at the wrong URL, or a sitemap a crawler can't reach all quietly limit how much of your site gets seen at all, regardless of how well any individual page is written. This tool checks the fundamentals alongside everything else.
Scoring
How is the score calculated?
Every check carries a weight from 1x to 5x, and heavier checks - like whether GPTBot or ClaudeBot can even reach the page - move the score far more than lighter ones, like meta description length. Only checks that reach a real pass, warn or fail count toward the total; inconclusive checks are left out instead of guessed at. A hard crawler block caps the whole score in the lowest band, because that finding matters more than the rest combined.
| Band | Score | What it means |
|---|---|---|
| Invisible | 0–39 | AI systems can barely find, read or cite this page as it stands. |
| Partial | 40–74 | Reachable and partly readable, with real gaps limiting how reliably it gets cited. |
| Solid | 75–100 | Crawlable, machine-readable and structured the way AI systems tend to extract from. |
Free vs. locked
What's free and what's locked?
Every check this tool runs, and what it found on your page, is free and shown in full - 100% of it, including the checks that failed. What's locked behind an email address is the fix: a priority-ordered list of what to change first, paste-ready code where that applies, and a corrected version of your page's structured data, not just a description of what's wrong with it.
Seeing your score and every finding costs nothing - no card, no email, no catch.
FAQs
Common questions
Is this the same as a Google PageSpeed or general SEO audit?
No. This tool checks AI-specific visibility: whether crawlers like GPTBot and ClaudeBot can reach a page, whether its structured data resolves cleanly, and whether the page is shaped so an AI system can extract and cite an answer from it. There's real overlap with technical SEO fundamentals, but it isn't a page-speed or ranking-factor tool - for that, try our free website audit instead.
Does a low score mean my site is broken?
Not necessarily. It means AI systems are less likely to reliably find, read or cite the page as it stands, which is a different problem from the page working fine for human visitors or ranking normally on Google. Plenty of functional sites score poorly here simply because nobody has looked at this specific thing yet.
Why couldn't you get a score for my site?
Usually one of two reasons: the site is blocking automated requests behind a bot-check or challenge page, or the page renders its content with JavaScript after load in a way our fetcher can't see yet. Both show up as a plain inconclusive result instead of a guess - the result itself tells you which one applies.
What happens with the email address I give you?
It unlocks the fixes for that specific report, delivered as a private link only you have. We don't add you to anything else unless you tick the separate marketing checkbox on the form, which is off by default. The report link itself is stored for 30 days - see our privacy policy for the rest.
How often should I run this?
After any change to robots.txt, structured data, or page templates, and every few months regardless - what the named AI crawlers respect changes fairly often, so a page that passed a check six months ago is worth rechecking.
Read more
Further reading
For the fuller picture of what drives AI citations beyond this tool's checklist - crawler access, content structure, schema and the trust signals that compound over time - seehow to get cited in AI search.
Looking for page speed, technical SEO hygiene or on-page fundamentals instead? Those aren't this tool's job - try our free, fully openwebsite audit tool.
If the list of fixes feels like a lot, we can just do it for you.
If the audit turns up more than you want to tackle alone, this is exactly the kind of technical SEO and AI-visibility work we do for clients, and asking about it costs you nothing.