How does Bot Appétit check whether AI can read my page?
We read your page the way a machine does, check it against 66 public standards and conventions, and tell you in plain English what a machine can and can't use. Here's every step, from the moment you press scan to the score, and what we deliberately leave out.
Who asks
- AnyonePage scans. Free, no signup.
- Signed inSite scans and Your sites, with a free account.
- Sent a linkReads the report. No account needed.
A page scan
-
You enter an address
Any public address. No account, no signup.
-
Our bot fetches it
The page, plus robots.txt, the sitemap, llms.txt and the other site-level files machines look for first.
BotAppetitBot · named and signed (RFC 9421 Web Bot Auth)
-
We read the raw HTML
The way most AI crawlers do. A real browser renders the page only when the raw HTML looks thin.
Cloudflare Browser Rendering, only when needed
-
Every check returns a result
- Pass
- Worth a look
- To fix
- Couldn't judge
-
The score
0–100 and a verdict, stamped with the scoring version that produced it.
Performance: Google PageSpeed, on its own dial
-
Public page history
The result belongs to the page. Every scan of the address adds to its history.
A site scan: the page scan is the atom
-
Discover pages
From the sitemap or, where there isn’t one, the links on the homepage.
-
Sample by template
Up to 50 pages, chosen to cover each kind of page the site has.
-
A page scan on each
Steps 02 to 05 above, in full, for every page we read. Performance is measured on the homepage.
-
One site score
Judged across every page we read, not an average of their scores.
Outside services we call
- The site being scannedIts page and its site-level files, fetched by our signed bot.
- Cloudflare Browser RenderingA real browser, only when the raw HTML looks thin.
- Cloudflare’s public DNS resolverThe DNS records we check, so those lookups never reach your server.
- Google PageSpeed InsightsCore Web Vitals, on their own dial. Never part of the score.
- Google Safe BrowsingPart of the security checks.
What we fetch
A page scan fetches the page you entered, plus the site-level files machines look for first: robots.txt, the sitemap, llms.txt, security.txt, the favicon and the /.well-known/ agent files, with one AI-discovery DNS lookup and one missing-page probe. Every request comes from one of our two named bots, and every request the scanner makes is signed. We read the raw HTML the way most AI crawlers do, and open the page in a real browser only when that HTML looks thin.
Everything about our botHow a check is judged
Every check that applies to your page returns one of four results. Every check says why it matters and what good looks like.
- Pass
A machine gets what it needs here.
- Worth a look
It works, but a machine gets less than it could.
- To fix
A machine can’t get what it needs here.
- Couldn't judge
We tried and couldn’t look. It never counts against the page.
A check that doesn't apply to your page is left out entirely: no row, and it isn't counted.
How the score works
Your score comes from the 7 scored categories. Advanced and Good practice are shown but never counted, and performance sits on its own dial. The verdict is a plain-English name for where the score lands, and it can be held down when AI crawlers are shut out, however healthy the rest is. We don't publish the exact weights: that calibration took hundreds of real sites to get right. What we always publish is that something changed.
- Chef's kiss Nothing to send back — AI engines can read you loud and clear.
- Cleans the plate A strong plate. A few rough edges, nothing that stops a machine reading you.
- Needs seasoning The bones are good, but answer engines are leaving hungry.
- Undercooked There's a lot on the menu bots can't taste yet — worth a proper pass.
- Sent back to the kitchen AI engines can barely get a bite of this one — time for a full recipe rework.
Site scans
A site scan finds your pages from your sitemap or, where there isn't one, from the links on your homepage. It reads up to 50 of them, chosen to cover each kind of page your site has, so every template gets looked at. Each page gets the same full page scan, with performance measured on the homepage. The site score is judged across every page we read, not an average of their scores.
History and scoring versions
Results belong to the page, so every address keeps a history anyone can see. Every result is stamped with the version of our scoring that produced it. When our scoring changed between two scans, the history still shows the difference, and says our scoring changed. Only two comparisons show no number: “Not compared — the pinned pages changed”, when a site scan's pinned pages differ between runs, and “Can't compare”, when the earlier version was never recorded.
What we changedWhat a scanner can't tell you
We measure whether a machine can read your page. We don't measure rankings, citations or what AI assistants say about you, and some things only a person can judge: whether the writing is good, whether the offer is clear, whether the page answers the question. That's where a human helps.
Want this handled? No pitch, just a plain-English chat.
Book a call