See what a crawler
actually receives.

Built in Denver by an agency that runs its own client sites through it — the demo report is one of ours.

The scan is free. Monitoring $19/mo · Agency reports $49/mo →

What an answer engine has to get through

Both diagrams are drawn from real readings in this dataset — not illustrations of an idea.

What a scan turns up

Three findings from sites already measured here. None of them appear in a validator, an uptime check, or a rank tracker.

476,540 bytes to deliver 2.3% text

The heaviest page measured through this checker. A crawler pays that cost on every URL, because inline CSS cannot be cached between pages. The site looked fine to every human who visited it.

HTTP 200 on a file no crawler could read

An origin bot shield answered /robots.txt with a verification page and a success status. Crawlers do not run the JavaScript — they read the interstitial as the file. Every uptime monitor called it healthy.

Two locations sharing one coordinate

Two business entities declaring the identical latitude and longitude. Perfectly valid schema; no validator has an opinion. It tells a retrieval system the two places are the same place.

What the report gives you

Two scores, because they fail independently

Crawl clarity — can a machine fetch, read and tell your pages apart. User clarity — can a reader tell what this is and who it belongs to. A site can be perfectly crawlable and say nothing coherent.

An AI visibility score, and its limits stated on the page

Whether an engine can reach you, read you, and quote a fact about you — three stages, in the order an engine hits them. It does not claim to know whether ChatGPT cites you; no server can measure that.

Every answer engine, one at a time

OAI-SearchBot indexes, ChatGPT-User fetches live, GPTBot trains. Three separate permissions a site routinely splits without meaning to. The report shows what each was allowed to do and what your edge actually gave it.

The ranked list of what to fix, with the points attached

Every component shows its weight, what was measured, and what the number would be if it were fixed. You can work top-down and stop when the remaining items stop being worth the hour.

What we have actually found

Each one measured on a real site. None of them show up in a rank tracker, a validator or an uptime check.

What you get, and what costs money

The paid rows are the ones that spend requests or run on a schedule. Everything a single fetch can answer stays free, and no row anywhere changes a measurement.

Two things worth paying for

The scan is free and stays free. What costs money is the deliverable and the watching — the two jobs a tool cannot do by being run once.

Nothing a licence unlocks changes a measurement. The moment a payment can move a result, every result here is worth less — including the ones you would be paying for.

Questions people ask before scanning

Does this change anything on my site?

No. Every request is a read — the same GET a crawler makes. Nothing is submitted, no form is filled, no page is written to, and nothing is installed.

Will my domain show up anywhere public?

No. Aggregate findings are published on the dataset page and no scanned domain is ever named there. One line in robots.txt keeps you out of the dataset entirely.

What if my site is already fine?

Then you will know in about a minute and pay nothing. That is the likeliest outcome for a well-built site, and it is worth confirming rather than assuming — the most common serious finding on this site so far was a file that returned 200 to every crawler while serving them a verification page instead of the content.

Is this just an SEO audit with new words on it?

No. An SEO audit reads your page as a browser. This sends the requests an answer engine sends — as GPTBot, ClaudeBot, PerplexityBot and the rest — and reports what your edge gave each one. A site can pass every SEO check and still refuse two of the four biggest answer engines.

Can I pay to improve my score?

No, and that is deliberate. Nothing a licence unlocks changes a measurement, a finding or a grade. The moment a payment can move a result, every result here is worth less — including the ones you would be paying for.

What it will not do

Report a defect from a failed measurement

If a scan is rate-limited, blocked, or the pages render client-side, the result says so and the affected components are excluded from the score. An unmeasured component is never counted as zero — a false clean bill of health is more dangerous than a false defect.

Claim to know a ranking algorithm

Where a threshold comes from published research, the finding names the study. Nothing here is presented as a formula for being cited, because no such formula is published.

Name your domain anywhere public

Aggregate findings are published at the dataset and no scanned domain is ever named in it. One line in robots.txt keeps you out of it entirely.