Can AI read your page?
Check whether ChatGPT, Claude, Perplexity and Google can crawl, read and quote a page. Dorqa reads it, tests it with five AI crawlers, and scores 24 checks from the 120-point AI Readiness Framework.
Free, the full report, no sign-up.
How Dorqa scores a page
The same method runs for every page, and every number in the result can be traced back to it.
- It reads the page. Dorqa fetches the address as an ordinary visitor, following up to 5 redirects and checking each one. It also reads the site’s robots.txt and llms.txt.
- It asks as five AI crawlers. The page is requested again with the user agents of GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and CCBot. A 401, 403, 429 or 503 counts as refused.
- It scores 24 checks. Each check carries its points from the 120-point framework. A pass earns full points, a partial earns some, a fail earns none. Author and date are scored on articles only.
- Score = points earned ÷ points measured × 100. At most 67 points can be measured on an article and 61 on any other page. The 53 points that need a person or a real browser are left out of both numbers, so a missing judgement neither lowers nor inflates the score.
- Grade. 90 and above is Ready, 75 to 89 Mostly ready, 50 to 74 Needs work, below 50 Not ready. Two rules override the number: if the page cannot be indexed or quoted (noindex, Googlebot blocked, not answering 200, snippets blocked), it is Not ready whatever it scores; if any AI crawler is refused, it is at best Needs work.
- AI crawler points are shared. The 3 points are split across the five crawlers. A crawler counts only if robots.txt allows it and the live request was not refused, so 3 of 5 earns 1.8.
- Fixes are ranked by points lost. A refused AI crawler always comes first, because robots.txt never shows it.
adschoolmaster.com: 52 of 61 points measured, so 52 ÷ 61 × 100 = 85, Mostly ready. Crawl and index 22/22, content organisation 11/15, technical structure 9/12, schema and AI access 10/12. All five AI crawlers got 200. Top fix: no question-and-answer block (3 points).
Results are kept for 24 hours per address, so checking the same page again shows the same result and does not use a check.
24 checks, and what each one is worth
The points come from the 120-point AI Readiness Framework. Only checks that can be measured from the page are scored. The score is points earned out of points measured.
Crawl and index eligibility
- robots.txt can be read 2 pts
A server error on robots.txt makes Google stop crawling the whole site. - Googlebot may crawl the page 5 pts
Google's AI Overviews and AI Mode use the same crawl. - The page answers 200 2 pts
After redirects, an ordinary request must get the page. - No noindex 6 pts
Meta robots tag and X-Robots-Tag header. - Snippets not blocked 5 pts
nosnippet or max-snippet:0 stops Google quoting the page. - Canonical points to the page 2 pts
Content organisation
- One H1 3 pts
- Headings do not skip levels 2 pts
- First paragraph of 80 words or fewer 4 pts
Short enough to be lifted as an answer. - A question-and-answer block 3 pts
FAQPage schema, expandable answers or question headings. - A table or list in the content 1 pts
- 3 or more internal links in the content 2 pts
Menus and footer are not counted.
Technical structure
- A <main> element 2 pts
- <article> or <section> elements 2 pts
- Title of 50 to 60 characters and a meta description 3 pts
- Images have alt text 2 pts
- Icon buttons and links have labels 1 pts
- Open Graph tags 2 pts
Schema and AI access
- JSON-LD that parses and matches the page 4 pts
Schema whose headline contradicts the title and H1 loses points. - A schema type that says what the page is 3 pts
- Author shown on the page 3 pts
Articles only. - Published or updated date shown 3 pts
Articles only. - AI crawlers can reach the page 3 pts
robots.txt and a live request as GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and CCBot. - llms.txt 2 pts
Optional. Small measured effect today.
robots.txt can say yes while your CDN says no
Most checkers read robots.txt and stop. But a firewall, CDN or bot-protection setting can refuse AI crawlers before robots.txt is ever read, and nothing in the file shows it. We found exactly this on a site of our own: robots.txt allowed every crawler, and GPTBot, ClaudeBot and CCBot were refused at the edge.
- Requests the page as GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and CCBot
- Reports each crawler's robots.txt verdict and its live answer side by side
- Puts any refusal at the top of the fixes, whatever else scores
The requests come from Dorqa's servers with each crawler's user agent. A site may treat the real crawler differently, because it can check the crawler's IP address.
A refusal here is a strong signal. Confirm it in your CDN's logs.
Content quality is 25 of the framework's 120 points: originality, intent and depth. Dorqa does not guess them, so they are listed as not measured and kept out of the score. So are fan-out coverage, statistics with sources, and agentic browsing. The same rule runs through all of Dorqa: how we handle data quality.
A high score removes the blockers. It does not earn the citation.
Whether ChatGPT or Google's AI Overview names your page depends on the content and on what else answers the question. AI Visibility shows which prompts mention you and which sources get cited instead. Content Intelligence shows what the ranking pages cover that you do not.
About the checker
24 checks from the 120-point AI Readiness Framework that can be measured from the page itself: whether search and AI crawlers may fetch it, whether it can be indexed and quoted, how the content is organised, the page structure, the schema, and whether AI crawlers actually reach it.
Because robots.txt is not the only gate. A firewall, CDN or bot-protection setting can refuse AI crawlers even when robots.txt allows them, and nothing in robots.txt shows it. Dorqa requests your page with each crawler's user agent and reports any refusal. Your site may treat the real crawler differently, by its IP address, so check your CDN's logs to confirm.
Content quality: whether the page is original, complete and specific. That needs a person, or a model, to judge, and Dorqa does not guess. It also leaves out query fan-out coverage, statistics with sources, and agentic browsing, which needs a real browser. These are listed with every result and left out of the score.
Points earned divided by points measured, times 100. Each of the 24 checks carries its points from the 120-point framework: full points for a pass, part for a partial, none for a fail. Author and date count only on articles, so the most a page can be measured on is 67 points, or 61 for a page that is not an article. The 53 points Dorqa cannot measure are left out of both numbers, so they neither lower nor raise the score.
Yes, the full report: every check with its evidence, and every fix. Up to 5 checks a day without an account, 30 with a free Dorqa account. A page checked in the last 24 hours shows the same result and does not use a check.
No. It means nothing on the page stops AI from reading and quoting it. Whether AI actually cites you depends on the content and on what else answers the question. Dorqa's AI Visibility tracks that.
Be first. Not just found.
Free tier, no card. See every check, then track whether AI actually cites you.