Why this matters now
People increasingly ask an assistant instead of typing a query, and assistants answer from two places: what they absorbed during training, and what they can fetch when asked. Both depend on access. A single line in a robots.txt file can remove your business from an AI's world entirely, and most site owners have never looked at which of these crawlers they are blocking, because the blocks are often added by a plugin, a host, or a well-meaning developer worried about scraping.
The second half is stranger. Assistants will describe a business confidently and get it wrong, and searching does not always fix it: it can make it worse by finding a similarly named company. Both things have happened to this site. An AI answer once attributed rank.fast to a founder who has no connection to it, invented from nothing. When we first ran this very tool on ourselves, one assistant described us as a parked domain for sale, another found a different company with a similar name, and a third got us exactly right. If that is what assistants are telling people about your business, you would rather know.
What the check does
| Crawler access | Reads your robots.txt and reports, one by one, whether GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended, meta-externalagent, CCBot and Bingbot are allowed to fetch your site. |
|---|---|
| Machine readability | Whether you publish an llms.txt, what structured data your homepage carries, and whether your content is readable without running JavaScript. |
| Memory versus search | What ChatGPT (GPT-5.6), Claude (Sonnet 5) and Grok (4.6) each say about you from training alone, and separately what they say with web search enabled, including the pages they cite. The gap matters: being unknown from memory but findable by search is a different problem from being neither. |
| Monthly watch | Optional. Give us an email address after a check and we run it again once a month, then write to you only if something moved: an assistant that stopped knowing you or started, a crawler you began blocking, structured data that changed. If nothing changes you hear nothing. |
| Conflicts | Where either assistant's answer, in either mode, contradicts your site's own description of itself. |
What it isn't
This asks three assistants, not every assistant, and it asks their underlying models with web search enabled rather than driving the consumer apps themselves, so wording will differ from what you would see in the app even when the substance matches. Model tier matters too, and not in the way you might expect: when we tested the same question on a more capable model, it declined to guess and said plainly that it could not find reliable information, while the mid-tier models confidently described a different company with a similar name. We deliberately test a mid-tier model, because that is what most people are talking to, and because a confident wrong answer is the outcome worth knowing about.
Different systems were trained on different data at different times and search returns different things hour to hour, so treat this as two informative data points rather than a verdict on the whole AI ecosystem. Blocking a crawler is also a legitimate choice, and plenty of publishers make it deliberately. The point is that it should be a choice you made, not one you inherited. Nothing here promises visibility in any assistant; nobody can.If none of them has heard of you
That is normal for newer and smaller sites and it is not a fault. This site was unknown to all three from memory when we built this tool, eleven days after launching. With search on, one described us accurately, one confused us with a similarly named company, and one reported us as a parked domain for sale. Same site, same moment, three different answers: that spread is exactly what this tool exists to show you. The levers are unglamorous: let the crawlers in, make your pages readable without JavaScript, state plainly who you are in structured data, publish things worth citing, and get indexed where these systems look. Our free technical audit covers the indexing side, the index checker tells you which pages Google holds, and we publish everything we measure about doing exactly this on a site that started with nothing.