Every finding, with the proof.
AI crawlers allowed in robots.txt
3 / 10 ptsA customer's ChatGPT is turned away before it sees a single page.
robots.txt: User-agent: ChatGPT-User / Disallow: /
Fix: Remove the ChatGPT-User block. Live browsing agents are customers, not scrapers.
Site serves a request from an agent
7 / 7 ptsYour bot protection lets an agent through.
GET / as an AI agent: 200 OK
Machine readable business facts
5 / 10 ptsYour business is described. Your prices and opening hours are not.
JSON-LD: LocalBusiness found; no Offer, no openingHoursSpecification
Fix: Add your opening hours and service prices as structured data.
WebMCP tools an agent can call
0 / 14 ptsAn agent cannot take a single named action here.
document.modelContext: no tools registered
Fix: Register check_availability and book_appointment. We draft them from your page.
Forms a machine can fill in
5 / 9 ptsHalf the fields on your contact form have no label.
4 of 8 fields have no label tied by id
Fix: Give every field a label and an autocomplete token.
No captcha blocking the last step
6 / 6 ptsNothing stops an agent at the final step.
No reCAPTCHA, hCaptcha or Turnstile found on the page
See what we found on your site, and how to fix it.
17 checks across four areas, from your robots.txt to whether a booking can be finished.
Give agents something to do.
WebMCP hands an agent named actions. We draft yours from your own page.
The new standard from Google and Microsoft, and the heaviest check we run.
Hand the fix list to your developer.
Most expensive problem first, written to hand over without a meeting.
Changed something? Press Rescan on your report and watch the number move.
Why not just ask ChatGPT about your site?
Your robots.txt, your bot wall, the raw page and a real browser, the way an agent meets them.
Every finding quotes what we saw, so you can check our work.
Reading your page is table stakes. Booking and buying on it carry the most weight.
No account. Your score is yours, with its own link to share.