Can AI actually read your page?
Paste any URL. We fetch the page, read its robots.txt and llms.txt, and run 31 deterministic checks against it — then show you every result, the actual values we found, and what to do about each one.
Free · no signup · no card · any public page, not just your own
Answers within 108 characters
No FAQPage markup found
6 Q&A pairs visible on the page
4 of 9 headings are questions
Ranking on Google and being quoted by AI are not the same job.
They overlap — both need a page a crawler can fetch and read. They come apart on what happens next. A search engine ranks your page and sends someone to it. An answer engine reads your page and rewrites it into a sentence, and whether you are named in that sentence depends on things most SEO audits never look at.
For applicant tracking compatibility, most people land on Rival A or Rival B. Yourbrand is worth a look if you want the score broken down by section rather than a single number.
It has to answer, near the top
A 353-character introduction before the page answers its own headline is fine for a reader who scrolls. For a model assembling a two-sentence answer, it is the difference between being quoted and being skipped.
It has to be in extractable shapes
Lists, tables and question-shaped headings survive being lifted out of context. A beautifully written wall of prose does not, no matter how good it is.
It has to say who you are
A model attaching a claim to a brand needs the page to state, in machine-readable terms, who published it. Most pages simply never do.
31 checks, in seven groups.
Seven groups rather than a flat list of 31 ticks, because seven decisions are something you can act on and 31 boxes are a spreadsheet. Each group carries its own share of the 100 points — and the shares are not equal, because the problems are not.
Can AI reach it
5 checks · 22 of 100 points
Whether an AI crawler can fetch this page at all — and whether you have accidentally told it not to.
- robots.txt rules for GPTBot, ClaudeBot, PerplexityBot and Google-Extended
- Whether the page is indexable, or carries a noindex it has forgotten about
- A canonical URL that points somewhere sensible
- Whether /llms.txt exists
- 01AI crawlers allowed (robots.txt)6
- 02Page is indexable5
- 03Server-rendered content5
- 04Canonical URL declared3
- 05Clean 200 response3
Nothing about this score is a judgement call.
Hover any segment. The widest are the ones that stop a page being used at all — a robots.txt rule blocking GPTBot outweighs a missing publish date several times over, and the bar says so rather than treating every check as equal.
Fixed weights, summing to 100
Each of the 31 checks carries a weight, and those weights total exactly 100. They do not move between runs, which is what makes two checks a month apart comparable. A test in our build fails if that total ever drifts.
Deterministic, not model-judged
No language model decides whether your page passes. Every check is a rule applied to your HTML. The same page always scores the same, so a change in the number means a change on the page — not a change in an AI's mood.
Unmeasurable checks are removed, not failed
If a check cannot apply to your page, it comes out of both the top and the bottom of the fraction. A blog post is not marked down for having no pricing schema. The alternative — failing you for it — would make the score meaningless across page types.
That is the whole thing. We publish it because a score you cannot reproduce is a score you have to take on trust, and this product's entire argument is that you should not have to.
One request, 15s timeout, identified as CoreCitedBot.
Which AI crawlers you allow, quoted line by line.
Present or not — no guessing either way.
Pure rules against your HTML. No model involved.
Impact first, then by how easy each one is.
A perfect score is not a promise.
These checks remove the technical reasons an AI assistant cannot use your page. They cannot make it want to. Being named in an answer also depends on what exists about you elsewhere — reviews, forum threads, documentation, coverage — and roughly a third of what gets cited in this category is a forum thread you do not control.
The bottom three are industry-wide gaps, not ours alone. Every vendor in this category has them; some show you a number anyway. We would rather you read it here than find out after the invoice.
The report is free. The scale is what you pay for.
- ✓One page, all 31 checks
- ✓The real score, nothing blurred
- ✓The actual values found on your page
- ✓A shareable report link
- ✓10 pages instead of one
- ✓PDF export of the report
- ✓Saved history — see what changed
- ✓20 AI visibility checks
- ✓50–5,000 pages
- ✓Site-wide findings: duplicates, orphans, chains
- ✓Weekly re-runs with tasks and owners
- ✓AI visibility across all six surfaces
We are deliberate about this: you can read the entire report without giving us anything. The account is for keeping it — the file, the history, the pages you did not check. Hiding the answer to force a signup would make this tool worthless to link to, and links are the only reason we can afford to give it away.
Questions about this checker
Is this really free, with no signup?+
Yes. Enter a URL and you get the complete report — all 31 checks, the real score, and the actual values found on your page. Nothing is blurred or held back. A free account adds a PDF export, saved history and up to 10 pages; a paid plan adds whole-site crawling and weekly re-runs.
What is the difference between this and a normal SEO audit?+
A normal SEO audit asks whether Google can rank your page. This asks whether an AI assistant can quote it. Those overlap — both care about crawlability and speed — but they diverge on what matters most. An AI answer needs a page that states its answer early, structures it in extractable pieces, and identifies who published it. A page can rank well on Google and still be almost impossible to quote.
How is the score calculated?+
31 checks, each with a fixed weight that sums to 100. The score is the weight you earned divided by the weight that could be measured on your page. It is entirely deterministic — no model decides anything — so the same HTML always produces the same number, and two checks a month apart are directly comparable.
Why do some checks say 'not measurable'?+
Because a check that cannot apply to your page is removed from the total rather than failed. A blog post is not marked down for having no pricing schema. Those checks are excluded from both the top and the bottom of the fraction, so they cannot quietly drag your score down.
Can I check a competitor's page?+
Yes. It works on any publicly reachable page. It reads only what a normal browser or crawler would — the HTML, robots.txt and llms.txt — so there is nothing private involved.
Can I share the report?+
Every report gets a permanent link that anyone can open without an account. Copy it, send it to a developer or a client, and they see exactly what you see.
Does passing every check mean AI will mention my brand?+
No, and anyone who tells you otherwise is selling something. These checks remove the technical reasons an AI assistant cannot use your page. Whether it chooses to name you also depends on what else exists about you across the web — reviews, forums, documentation, coverage — most of which is not on your site at all.
Do you store the page or my data?+
We store the result so the share link keeps working, along with the URL you checked. No account, no email, no page content beyond the check results. The IP is truncated before storage and is only used to stop one source hammering the tool.
Now find out whether AI actually names you.
A readable page is the floor, not the goal. The next question is whether ChatGPT, Claude, Perplexity and Gemini mention your brand when someone asks about your category.