Can AI Find Your Rental?

What the Checker Reads

A guest sees your finished page. An AI crawler often sees something much thinner: the code your server sends before any script runs. The checker requests up to five of your public pages that way and scores five things out of 100. Every line in your report quotes what it found on that run.

Why These Checks Matter

A blocked crawler means no citation from that assistant. OpenAI's crawler page is blunt about it: opt out of OAI-SearchBot and ChatGPT search leaves your site out of its answers, apart from plain navigational links. Anthropic warns that turning off Claude-SearchBot can cost you visibility in Claude's results. Perplexity tells owners who want to be listed to let PerplexityBot through.

Training crawlers are a separate choice. OpenAI runs GPTBot, its training crawler, under a separate setting from OAI-SearchBot. Perplexity says its search crawler doesn't feed model training at all. So your report lists the training crawlers without taking points off when they're blocked. That decision is yours.

A firewall can overrule robots.txt. Perplexity's docs include a walkthrough for letting its crawlers past a web application firewall, which tells you how often this goes wrong. Our test borrows each crawler's name but not its network address. A refusal means some rule reacts to that name, so have a look at your host's bot settings.

Words behind JavaScript can go unread. Vercel and MERJ, watching real crawler traffic in late 2024, saw ChatGPT's and Claude's crawlers download JavaScript files and never run them. RESONEO's tests in January 2026 had ChatGPT, Claude and Gemini reading static HTML and missing anything a script added, with Bing Copilot as the one that did run scripts. So a booking widget or a review widget can be a blank space to most of these readers.

Google is the exception worth knowing. For AI Overviews and AI Mode, Google's guidance is that ordinary SEO still applies and that no extra files or special schema are needed. Googlebot runs JavaScript too. We keep schema in the score anyway, since it spells out your facts for every other machine reading your site. And llms.txt earns just 5 points because we haven't seen a major AI search engine say it reads that file.

What the Score Does Not Measure

The score tells you what a machine can read on your site today. It does not tell you whether ChatGPT names you today, and no outside tool can promise that it will. A high score means you have removed the obstacles. Getting recommended also rests on whether your pages answer real questions and on what other sites say about you, and this tool measures neither.

Answers also move around. Our July 2026 study put each question to Google's AI five times over, and any two runs had only about 47% of their cited companies in common. So this page scores the things you control, not one answer from one afternoon.

The same study is why owning a readable site matters. Over 450 destination questions, local and independent operators collected about 82% of citations. Airbnb and Vrbo together turned up in 17.1% of the answers.

Checking an Airbnb or Vrbo Link

Paste a listing and we read the listing page and that platform's robots.txt, because those are what a crawler would meet. The rules are the platform's, so the report says what you can and can't change. When we wrote this on November 5, 2026, Airbnb's robots.txt stopped GPTBot and ClaudeBot from the whole site and stopped PerplexityBot from /rooms/ pages, where listings live. It allowed OAI-SearchBot and Bingbot. Vrbo's file allowed all five search crawlers on the listing we tested.

A listing can score well and still leave you with nothing of your own: a guest who follows that citation lands on the platform, in the platform's checkout. If you also have a website, check that next. It is the one place where every point on this page is yours to win.

Checker Questions

Does a high score mean ChatGPT will recommend me?

No. It means crawlers can get in and can read your facts, which is the part you control. Recommendations also depend on your content and on what other sites say about you.

What does the checker do to my site?

It only reads. It requests up to five public pages, plus robots.txt, llms.txt and your sitemap, the same way any visitor could. It asks for the page three more times under crawler names to test your firewall. It logs in nowhere and changes nothing.

Is it free?

Yes. Your score and your two biggest gaps cost nothing and need no signup. The whole report lists every check, what we found and how to fix each gap. It opens when you give your name, email and phone, and someone from Cavmir may contact you about it.

Why did I get no score?

Your page refused us both as a browser and as a crawler. Our request comes from a shared cloud network, which some firewalls block on sight, so the real crawlers might do better or might not. We won't guess at a number. Open the full report and a person at Cavmir can read your site by hand.

Should I block AI crawlers?

Keep the search crawlers open if answers matter to you. Training crawlers are a separate decision, and you can shut them out without touching the others. We let both kinds onto cavmir.com, since we'd like our research to get cited.

Can I share the result?

Yes. The full report comes with a private link and a print button you can use to save a PDF, so you can pass it to whoever runs your website. The link shows the audit only, never your contact details.

Can Cavmir fix what it finds?

Yes, if you want that. It is part of our generative engine optimization (GEO) service. Your web person can also work straight from the fix list. If you want to learn the subject first, start with our long GEO guide for rental owners.

Where to Go Next

Sources

1. OpenAI's page on its crawlers. developers.openai.com/api/docs/bots
2. Anthropic's help article on its crawlers. support.claude.com
3. Perplexity's page on its crawlers and firewalls. docs.perplexity.ai/guides/bots
4. Google's Search Central page about its AI features. developers.google.com
5. Vercel and MERJ's crawler research from December 2024. vercel.com
6. RESONEO's crawler tests from January 2026. think.resoneo.com
7. The llms.txt proposal by Jeremy Howard. llmstxt.org
8. Airbnb and Vrbo robots.txt files, read November 5, 2026. airbnb.com/robots.txt and vrbo.com/robots.txt
9. Cavmir's own study from July 2026.