- Dwg
- VTA-T1
- Title
- AI readiness check
Dwg VTA-T1Free · no sign-up
Can AI assistants read your site at all?
llms.txt, AI-crawler access, structured data: what the bots actually see when they fetch your homepage. Thirty seconds. No account.
Six checks
Readiness is whether you can be cited.
We fetch your homepage, robots.txt and llms.txt once, as ourselves, and run the same six checks on every site. No LLM involved. Deterministic. Nothing stored.
Readiness tells you if you can be read; the Snapshot tells you whether you're actually cited. Run one — $121.
What each check means
| Check | Pass | Why it matters |
|---|---|---|
| AI-crawler access | robots.txt allows the answer-time crawlers (OAI-SearchBot, Claude-SearchBot, PerplexityBot, ChatGPT-User, Claude-User, Perplexity-User) | Blocked crawlers cannot fetch the page a citation would point to. Blocking training crawlers only (GPTBot, ClaudeBot, Google-Extended) is a legitimate choice and shows as a review item. |
| Homepage response | A 200 with text/html within a few seconds | Crawlers have short patience budgets. |
| Visible page content | 600+ characters of text in the HTML before JavaScript runs | Most AI crawlers do not execute JavaScript. A client-rendered app is an empty shell to them. |
| Structured data | JSON-LD naming who you are (Organization, LocalBusiness, SoftwareApplication…) | The cheapest way to state your identity unambiguously. |
| llms.txt | A plain-text /llms.txt | Optional. A concise one tells assistants what you do in your own words. |
| Indexing signals | A canonical link, no noindex | Citations point at one URL; noindex tells everything to stay away. |
Method details are on the methodology page.
Why readiness comes first
If an AI assistant can't read your site, it can't cite it. That's the whole argument, and it's the reason the readiness check exists.
Think about what happens when a buyer asks a question with web search on. The engine fetches pages, reads what it can, and builds an answer from the sources it managed to parse. A page that blocks the crawler, returns an error, or loads as an empty shell never enters that process. You don't lose the comparison; you're not in it.
So the six checks are about access, not about persuasion. The tool fetches your homepage, robots.txt and llms.txt once, as VisibilityToAI-ReadinessCheck, and runs six deterministic checks. No LLM is involved, which means the result is the same every time you run it. Also, nothing gets stored and there's no account.
Readiness is the precondition. Nothing more, nothing less.
Fixing a failed check
Each check has a practical fix, and most are smaller than you'd expect.
AI-crawler access. Open your robots.txt and look for rules that block OAI-SearchBot, Claude-SearchBot, PerplexityBot, ChatGPT-User, Claude-User or Perplexity-User. Those are the answer-time crawlers, the ones that fetch pages when someone asks a question. Blocking only the training crawlers (GPTBot, ClaudeBot, Google-Extended) is a legitimate choice, and the tool shows it as a review item rather than a failure. Decide on purpose.
Homepage response. You want a 200 with text/html, returned within a few seconds. If you get redirects to a login wall, a 403 from a bot filter, or a slow origin, fix that first.
Visible page content. This one catches a lot of modern sites. The check looks for 600+ characters of text in the HTML before JavaScript runs, because most AI crawlers don't execute JavaScript. If your homepage is an empty div that a script fills in, the crawler sees nothing. Server-side rendering or static output solves it.
Structured data. Add JSON-LD that says who you are. Organization, LocalBusiness or SoftwareApplication, whichever fits your business.
llms.txt. A plain-text file at /llms.txt. It's optional, so treat a miss as a nice-to-have, not a fault.
Indexing signals. Check for a canonical link and make sure there's no noindex left over from a staging build. That one gets forgotten more than most.
Fix them in order of how much they block. Crawler access and page content come first, because everything else depends on them.
Readiness is not visibility
Passing all six checks doesn't mean you're recommended.
Readiness tells you the door is open. It says nothing about whether anyone walks through it. An engine can read your homepage perfectly and still name three competitors when a buyer asks who to call. The readiness check can't see that, because it never asks an engine a question.
That's a different measurement. Visibility is whether AI assistants name and cite a company for the questions its buyers ask. The Snapshot measures it: 15 prompts, four engines (ChatGPT, Claude, Gemini and Perplexity), so 60 answers. Each answer is recorded for two separate things: Mentioned (your brand named in the prose) and Cited (one of your own domains in the source list). They're never blended into one score.
The verdicts are where it gets useful. A vertical can come back as Recognised, not cited, which means you're named but rarely cited, and that points to a page problem. It can come back Entrenched, where one rival appears in the majority of answers. Or Open, where no vendor is named in more than about a quarter of answers.
A sample isn't a census, though. The same question tomorrow can get a different answer, and a citation is a source link, not a click. So run the readiness check first, fix what fails, then find out whether the engines actually name you.