Can the answer engines reach your pages?
Before any score matters, there is a simpler question: when ChatGPT, Claude, or Perplexity go looking for an answer, can they actually fetch your pages — and is your content in the HTML they read? Access is the first gate. If an engine can't reach a page, nothing downstream — readiness, citation — can happen, however good the page is.
I think of your AI presence as a chain. Access is the first link: the mouth of the pipe. A page has to be reachable and readable before it can be judged ready, and ready before it can be cited. A break at the first link stops everything after it.
Access → readiness → citation. The first link gates the rest.
So I run two checks: can each citation crawler fetch your site, and does your server send them the words?
I read your site's robots.txt and check whether each of these crawlers is allowed to fetch your site root:
These are citation crawlers — the bots that fetch pages to answer a user's question in the moment. They are not the same as training crawlers (GPTBot, ClaudeBot, CCBot, Google-Extended), which collect text to train models. I deliberately track only the citation crawlers, because blocking a training bot does not change whether an engine cites you — but blocking a citation bot does, directly.
When your robots.txt disallows one of these from your site root, I recompute the blocked set on every crawl and that chip turns red:
A blocked citation bot forfeits that engine's citations entirely. No matter how strong the page, the engine can't fetch it to quote it.
Each crawler you block is one answer engine that can never cite you. A single over-broad rule in your robots.txt can quietly remove you from an engine's answers — the page still exists, but to that engine it may as well not.
Being allowed in is only half of access. Citation crawlers don't run JavaScript — they read the HTML your server sends, and nothing more. If your words are drawn into the page by script after it loads, the crawler arrives to an empty room.
I compare the text in your raw HTML against the text in the fully rendered page.
I call a page JavaScript-only when its server HTML is missing about half or more of the readable text that's present once the page has rendered in a browser. Thin pages fall below a small floor and are exempt, so a genuinely short page doesn't trip the check by accident.
A JS-only page is effectively invisible to the crawlers that cite you. The fix is server-side rendering: make sure the words arrive in the HTML your server returns, not only after script runs in the browser. This is the same extractability concern that readiness measures, seen from the gate.
The allowlist and the raw-vs-rendered comparison both come from crawling your site — that's where I fetch your robots.txt and each page. Once a page is reachable and its content is in the HTML, the next gate is readiness: whether the page is shaped to be cited.
You don't have to hunt for problems. A blocked citation crawler or a JS-only page becomes a ranked item on your do-this-next list, so the first gate is the first thing I tell you to fix.