Does AI know
you exist?
ChatGPT, Claude, Gemini, and Perplexity each send a different crawler. Most sites block at least one by accident. Run the check, see who can read you and who cannot, and get a list of fixes ranked by impact.
We check these crawlers, by name
GPTBot
ChatGPT training
OAI-SearchBot
ChatGPT Search
ClaudeBot
Claude training
Claude-Web
Claude browsing
Google-Extended
Gemini + AI Overviews
PerplexityBot
Perplexity citations
Applebot-Extended
Apple Intelligence
CCBot
Llama, Mistral training
Bytespider
TikTok Doubao AI
Plus ChatGPT-User, anthropic-ai, Perplexity-User, Meta-ExternalAgent, cohere-ai, Amazonbot.
What the report covers
Six categories, every one of them actionable.
15 AI crawlers, named individually
Most "AI SEO" tools tell you whether you allow AI crawlers as a group. We show you each one: GPTBot, ClaudeBot, Google-Extended, PerplexityBot, Bytespider, the Apple bot, Common Crawl, and seven more, with the exact line in your robots.txt that lets them in or shuts them out.
Identity signals via Wikipedia, Wikidata, and Common Crawl
LLMs train on Wikipedia, Wikidata, and Common Crawl. If your brand is missing from all three, no model will mention you no matter how well you optimise. We check all three for free and tell you which are blank.
Structured data depth, not just presence
Organization, WebSite, BreadcrumbList, Article, Person, FAQPage, Service. AI engines parse each of these for a different purpose. We tell you which you have and which are missing.
Crawler-readable rendering
Most AI crawlers do not execute JavaScript. If your content only appears after React mounts, it is invisible to them. We measure how much of your page is in the server HTML versus how much needs a browser to see.
llms.txt and the LLM-first file family
The emerging conventions: llms.txt, llms-full.txt, sitemap.xml, robots.txt. We check each, tell you what is missing, and explain what to add.
Content citability patterns
Headings ending in question marks, definition-style sentences, lists, author bylines, dated content. AI engines preferentially cite content with these patterns. We score yours against them.
Methodology
The bit other tools hide.
We fetch your URL once, the same way an AI crawler would: regular Node fetch, normal user-agent, follow redirects. We pull your robots.txt. We HEAD-check your llms.txt, llms-full.txt, and sitemap.xml. We parse all JSON-LD on the page. We measure the words actually in your server HTML versus what would need a browser to render.
Then we ask three free APIs about your brand: Wikipedia, Wikidata, and Common Crawl. These three feed every major LLM at some point. If you are missing from all of them, no engine knows you exist regardless of what your site looks like.
We do not query ChatGPT or Claude themselves. Their APIs cost money and their web UIs are not scrapable at scale. Any free tool that claims to is either lying or using a model that pretends to be ChatGPT. What we check is upstream of those answers, the inputs that determine what those models can say about you.
Frequently asked
Straight answers.
Is this actually free?▼
Yes, no signup, no email gate, no card. We rate-limit to 5 reports per minute per visitor and cache results for 6 hours, so the same URL twice in a row returns instantly.
Are you really querying ChatGPT and Claude?▼
No, and any tool that says it does for free is misleading. ChatGPT and Claude APIs cost money, and their web UIs cannot be scraped at scale without breaking terms. What we do instead is check whether their crawlers are allowed to read you, and whether the upstream training data sources (Wikipedia, Wikidata, Common Crawl) know about you. Those are the inputs that determine whether the models will mention you.
What is the AI Visibility Score actually measuring?▼
A weighted average across six categories: AI crawler access (25%), structured data (20%), content citability (15%), identity signals (15%), crawler renderability (15%), and LLM-discoverability files like llms.txt (10%). Each category is scored from the individual checks underneath. The full breakdown sits below the score on every report.
Why does the report mention so many crawlers I have never heard of?▼
Because each LLM has its own crawler with its own name. GPTBot for ChatGPT training, OAI-SearchBot for ChatGPT Search, ClaudeBot for Claude training, Google-Extended for Gemini, PerplexityBot, Bytespider for TikTok, CCBot for Common Crawl (which feeds Llama and Mistral). If your robots.txt blocks a wildcard or names a bot incorrectly, you might be invisible to one engine and fully indexed in another.
My site got a low score. What now?▼
Read the "first five things to do" block at the top of your report. Those are ranked by impact. Most low-scoring sites recover quickly because the highest-impact fixes are usually trivial: add Organization JSON-LD, allow GPTBot in robots.txt, make sure server HTML contains the actual content. If you want help, hit our AIO service page, that is the engagement we run for this work.
How is this different from your free SEO audit?▼
The SEO audit looks at how Google ranks you in traditional search. This looks at how LLMs cite you in AI search. Different surface, different signals, different fixes. Many of the recommendations overlap (structured data, semantic HTML), but the crawler matrix and identity-signal sections are specific to this tool.
When you want help
Get found by the engines you actually care about.
The report tells you what to fix. If you would rather hand it over, our AIO team runs the implementation, monitors the score over time, and reports back monthly.
See AIO service