Questions about Bot Appétit

AEO (answer-engine optimisation) is making your site easy for AI answer engines — ChatGPT, Perplexity, Google AI Overviews — to read, understand, and cite. SEO is about ranking in traditional search results. Bot Appétit checks both, since they increasingly overlap.

Yes, completely free. A page scan needs no signup: enter an address and you get a 0–100 score with a category-by-category breakdown in about 20 seconds (the live performance numbers take a little longer). Site scans are free too. A site scan asks for an account only because it runs in the background for several minutes and need somewhere to keep your results.

There’s no catch. Every check, every score and every report is free, and a page scan needs no account at all. You don’t even have to tell us who you are. An account exists for one reason: a site scan runs in the background for several minutes, so it needs somewhere to put the results when it finishes. That somewhere becomes your history, so you can come back later and watch a site’s score move as things change. We’re Milk Moon Studio, a web studio. We built this because sites machines can’t read are the problem people hire us to solve, and the scan is how we show our work.

49 checks across AEO / content, AI access, AI signals, Structured data, SEO, Links and Security, plus 11 in Advanced (agent and MCP readiness) and 6 in Good practice, which we measure without scoring: 66 checks in all. On top of those, Core Web Vitals are measured live through Google PageSpeed. Every check is listed and explained on the “Everything we check” page.

PageSpeed tells you whether a page is fast. Bot Appétit tells you whether a machine can read it: AI access, structured data, content, SEO, links and security. We show PageSpeed’s numbers too, on their own dial, because speed isn’t the same question.

Because we tried and couldn’t look: the site refused that request, it timed out, or the page didn’t load fully. We’d rather say so than guess, so a check we couldn’t judge is left out of the score entirely.

Either the page changed or our scoring did. Every scan records the version of our scoring that produced it, so when ours changed between two scans, the history still shows the difference and adds “our scoring changed”, with a link to what we changed.

Because we couldn’t read it. When a site refuses our scanner, asks for a sign-in or doesn’t answer, we say what happened and give no score, never a zero. Where a firewall or bot protection turned us away, the report shows how to let our bot in, and you can scan again once it’s in place.

If your robots.txt blocks crawlers like GPTBot, ClaudeBot or PerplexityBot, those answer engines can’t read your content, so they have nothing of yours to quote. Bot Appétit flags it, and it holds the verdict down however good the rest of the page is: a site AI can’t read isn’t ready for AI, however tidy its markup.

Yes. Every scan of the same page is kept, so the results show a history of that exact address with the score change between each one. It includes scans anyone ran of that address, not only yours. Scan a page again after a change and you’ll see whether the number moved. There’s no signup and nothing to set up: the history builds itself.

Yes, if you ask it to. A page scan reads the one page you enter, along with that site’s robots.txt, sitemap, llms.txt, security.txt, favicon, its /.well-known/ agent files, an AI-discovery DNS lookup and a missing-page probe. A site scan works through a representative sample of your site instead, and gives every page it reads its own full report: the same checks in the same detail, because a site scan is simply many page scans. A site scan is free and needs an account; a page scan never does.

Because it isn’t an average of them. We judge each check across every page we read, then score the site the way we score a page. A problem that turns up on most of your pages counts against the site even when each page scores well on its own, and a few things that belong to the whole site, such as llms.txt and security.txt, only need to be right once.

Yes. Every scan gets its own permanent link you can copy, bookmark, or send to a client or developer. It opens the full report for anyone with the link, no signup needed, and it always shows that scan exactly as it was on the day, including the performance measured at the time. Every scan in the history has its own link too, so you can share a past result as easily as the latest one. A site scan can be shared the same way. A site scan is private to your account until you create a share link, and then anyone with that link can read the whole report, every page and every finding, without an account. That matters because the person who has to fix the findings is usually not the person who ran the scan. You can stop sharing at any time, and the old link stops working straight away.

We keep a record of each scan: the address, what we found and when. Results belong to the page, so anyone who scans the same address sees the same history. We don’t store your IP address. Now and then we keep a copy of a page to test our own checks against, and we never publish it. The privacy policy has the detail.

Not as a rule. We store what we found, not the page, and most scans are never copied. Now and then we keep a copy of a page to test our own checks against. It’s never published, shared or made searchable, and if you’d like the copies of your site removed, email jakes@milkmoonstudio.com and we’ll delete every one.

A page scan belongs to the page: anyone with its link can open it, and anyone who scans the same address sees the same history. A site scan’s list of pages is private to your account until you create a share link, though each page’s own result stays public like any page scan. The list of other pages you scanned on a site is only ever shown to you, in your own browser.

Yes, and it builds itself. Scan a second page on a site you’ve scanned before and the results show a list headed “Other pages you scanned here”, each linking back to its own report. The list is tied to your browser rather than to an account, through the analytics cookie our site sets. So it appears only in the same browser, and only when that cookie is there. Block it, or use a private window, and there’s nothing for us to match on. The list is never shown to anyone else. That’s deliberate: publishing a list of pages scanned on a domain would expose staging sites and unlinked pages that were never meant to be found, so the list stays yours.

Not a roaming one. BotAppetitBot doesn’t roam the web, keeps no index, and never goes looking for sites on its own. It fetches only when a person asks: one page for a page scan, or a sample of up to 50 pages on one site for a site scan.

Every request the scanner makes carries a signature you can check against our published key (RFC 9421 Web Bot Auth). The renderer, which loads a page in a real browser only when its HTML looks thin, is signed by Cloudflare rather than by us. Our bot page lists both user agents and where the key lives.