Twelve checks, each worth a set number of points. Your score is the total of the checks you pass. Here's every one of them, and exactly what we read.
Check my siteTwelve checks. One score.
robots.txt tells bots which pages they may read. If it blocks the bots ChatGPT, Perplexity, Claude, or Google use to look things up, your pages can't show up in their answers.
Some hosts and firewalls, including Cloudflare's AI bot blocking, refuse AI bots before robots.txt ever comes into play. Your robots.txt can say yes while your firewall says no.
AI tools and browsers treat sites without working HTTPS as less trustworthy, and some won't follow links to them at all.
Most AI crawlers read the raw HTML and don't run JavaScript. If your homepage only fills in after scripts load, those crawlers see a nearly empty page.
Guests ask AI specific questions: how many it sleeps, whether dogs are allowed, how far it is to the beach. AI tools can only answer from facts written on the page in plain text, not facts locked inside a booking widget or a photo.
schema.org has a VacationRental type that states what you rent, where it is, and how many it sleeps in a format machines read directly. It's the clearest way to tell an AI tool that this is a rental and here are the facts.
AI tools lean on reviews when they recommend a place to stay. If your only reviews live on the OTAs, those listings get the credit. Reviews on your own pages, with Review or AggregateRating markup, give AI something of yours to quote.
AI tools match your website to your Google Business Profile and OTA listings by name, address, and phone. Stating them in structured data on your homepage makes that match easy. If the details differ across those places, AI may treat them as different businesses.
Guides to your area and answers to common questions give AI tools a reason to treat you as the local expert, and they match the kinds of questions guests ask before they book.
A sitemap lists every page on your site, so crawlers can find each property page and not just the homepage.
The title and meta description are often the first summary of your site an AI tool reads. Open Graph tags control how your link looks when it's shared.
llms.txt is a proposed standard: a short plain-text file that tells AI tools what your site is and links to the pages that matter. AI companies haven't committed to reading it yet, so it's worth only 5 points, but it's quick to add.
How the score works
We don't curve it, and it doesn't change based on who built your site. Pass the 15-point robots.txt check and you get 15 points. Fail it and you get zero.
The pages and files AI crawlers look for first. We don't log in, crawl every page, or change anything.
Questions
Your homepage, robots.txt, llms.txt, your sitemap, and up to three property pages we find through your sitemap and homepage links. We also request your homepage the way two AI bots identify themselves, to see whether a firewall turns them away. We don't crawl the whole site.
No. We separate the bots that look things up to answer questions (OAI-SearchBot, ChatGPT-User, PerplexityBot, Claude-SearchBot) from the ones that collect training data (GPTBot, ClaudeBot, Google-Extended). Block the training bots if you like. Only blocking the answer bots costs points.
Usually. Most fixes are a robots.txt edit, a block of structured data, or text added around the booking widget. The fix guide spells out each one so you can send it to your booking engine's support team or your web person.
Some AI assistants can browse websites and fill in forms for a person, but most guests use AI to research and then book themselves. Either way, the AI has to be able to read your site first, and that's what this checks.
Not yet. We check that your name, address, and phone are stated clearly on your own site. Matching them against your Google Business Profile and OTA listings is worth doing by hand, and the fix guide tells you what to compare.
A short email when something in AI search actually changes for property managers, and what it means for your bookings. Unsubscribe in one click.