Free tool · AI visibility
Free AI citation checker
To check whether an answer engine can cite a page, two things have to be true: its crawler must be allowed to fetch that exact path, and the page must contain a passage worth quoting. This checks both, per engine, for the URL you enter. It cannot tell you whether an engine has actually cited you, because the consumer apps have no API and no tool can query them.
Free and unlimited. The per-engine audit needs no account; the live category run needs a free sign-in and is limited to one run a month, because it costs us money.
AEO depth layer
See the Live category run
The Live category run is computed from your own page and costs nothing to run. We ask for a free account because the expensive tools on this shelf stay free, and because it lets us tell you when your result changes. No card, no trial clock.
- We take the question your page is trying to answer, read from its own markup, and put it to an engine.
- The result says whether your brand is named in the answer, and which other names came up instead.
- It is a model API rather than the consumer ChatGPT or Perplexity app, which have no API and cannot be queried by any tool. The result says so on every path.
Signing in saves the details you entered above so we can follow up about your results. See our privacy policy for what we keep and why. The AI Citation Checker result you already have stays free either way. Privacy policy.
The engines, the bots they use, and what each one needs
Every row cites the vendor documentation it comes from, because bot tokens and behaviour change and are repeated wrongly across the SEO web. Training crawlers are deliberately absent: blocking them is a separate decision about model training and does not affect whether these engines can cite you.
| Engine | Bot that decides citation | How it selects sources | Source |
|---|---|---|---|
| ChatGPT search | OAI-SearchBot | Surfaces pages through its own search index. A site that blocks OAI-SearchBot is not shown in ChatGPT search answers, whatever else is on the page. | Vendor docs |
| Claude search | Claude-SearchBot | Indexes pages for search quality and retrieves them live when a user asks something. Blocking the search bot reduces both visibility and accuracy. | Vendor docs |
| Perplexity | PerplexityBot | Crawls and cites sources inline, with a visible link per claim. Of the engines here it is the most citation-forward, so a quotable passage matters most. | Vendor docs |
| Google AI Overviews | Googlebot | Built on Google's ordinary search index rather than a separate crawl, so anything that keeps a page out of that index keeps it out of AI Overviews too. | Vendor docs |
What a citation checker can and cannot know
This tool is named for a question it partly cannot answer, so the line is drawn here rather than left for you to discover.
| Question | Answerable? | Where |
|---|---|---|
| Can this engine fetch this exact path? | Yes, definitively | This tool, free |
| Is there a passage here worth quoting? | Yes, structurally | This tool, free |
| Does a model name my brand for my category question? | Yes, for a model API | The signed-in run here |
| Did the ChatGPT app cite my URL for a user? | No, and no tool can | Nothing. The apps have no API. |
| Which questions are my buyers actually asking? | No | Answer Radar |
| Am I named across five engines, tracked over time? | No | Answer Radar |
If every engine can reach the page and none would quote it, the fix is the content rather than the configuration. The AEO readiness score grades that in detail, the answer preview shows exactly what an engine could assemble from the page today, and the AI crawlability checker reviews the whole site's robots.txt rather than one path.
How to check whether AI engines can cite a page
Step 1
Enter the page you want cited
A page that answers a specific question. The check is path-level, so /guides/pricing and your homepage can and often do get different verdicts.
Step 2
Read reachable before citable
If an engine cannot reach the page, nothing about the content matters to that engine yet. That is a robots.txt fix and it takes a minute.
Step 3
If everything is reachable and nothing is citable, the page is the problem
That is the more common outcome and the more useful one. Permission is easy; having a passage worth quoting is the actual work.
Step 4
Then run the live category question
Sign in and the deeper layer takes the question your page answers, puts it to an engine, and reports whether your brand comes up at all.
When you outgrow this tool
One question, once, is a spot check. Answer Radar puts your buyers' real questions to five engines on a schedule and tracks whether the answer changes after you publish.
Checking one page is free and unlimited. Linkeddit is what you use when the question becomes which of your buyers' questions you are named in at all, across five engines, tracked over time.
FAQ
AI Citation Checker questions
What is checkable about citation from outside an engine, what is not, and why the training crawlers are deliberately left out of the verdict.
Can you tell me whether ChatGPT has actually cited my page?
No. Nothing can, and that includes every tool that says otherwise. The consumer ChatGPT, Claude and Perplexity apps have no API, and scripting the web apps is against their terms. What this tool does instead is answer the two questions that are genuinely checkable: whether each engine's crawler is permitted to fetch that exact path, and whether the page contains anything an engine would quote. The signed-in layer additionally asks a model API your page's own question, which is a real signal about your category and is not the same thing as the consumer app.
How is the signed-in run different from the ChatGPT app?
It is a model API. The app and the API are not the same product: the app runs its own retrieval, its own index, its own system prompt and its own personalisation, and their answers differ. We label the result accordingly rather than showing you an API answer and implying it is what a buyer sees in the app. If a tool shows you a ChatGPT-branded screenshot of a citation check, that is either an API call dressed up or a person taking screenshots.
Why does this ignore GPTBot and ClaudeBot?
Because they are training crawlers, not search crawlers, and blocking them is a deliberate and reasonable choice many publishers make about model training. It does not stop ChatGPT search or Claude search from citing you: those use OAI-SearchBot and Claude-SearchBot. Tools that lump them together produce an alarming red result for a site that made a considered decision, which is scaremongering rather than analysis.
How is this different from your AI crawlability checker?
That tool reads your robots.txt at the site root and reports a verdict for every AI bot, which is the right shape for a site-wide configuration review. This one reads a single path, groups the bots by engine, and pairs permission with whether the page has anything quotable on it. If you want to know whether your site is set up correctly, use that one. If you want to know why one specific page is not getting cited, use this one.
My page is reachable everywhere but citable nowhere. What now?
That is the most common result and it is good news, because it is the half you control. It means every engine is allowed in and none of them would find a self-contained passage worth lifting. The usual cause is an opening paragraph that is too long to quote whole, or a page that never states its answer in one place. The AEO readiness score grades that in detail and the answer preview shows you exactly what an engine could assemble today.
What is in the layer that needs an account?
One live run. We derive the question your page is trying to answer from its own markup, put it to an engine, and report whether your brand is named and which other names came up. The other-names extraction is heuristic, pulled from the prose, and it is labelled as what came up rather than as a verified competitor list. It costs us money, so it is one run per account per calendar month.
Why is the free part unlimited?
Because it costs us two cheap fetches of your own site and nothing else. Only the live run spends anything, so only the live run is metered. Putting a limit on the whole tool to make the limit feel like a feature would have been the easier product decision and the worse one.
More free tools
Every tool on the shelf is free, with no card and no trial. Most need no account at all. See all free tools.
Build the JSON-LD properties that make a page attributable rather than eligible for a rich result, with the reason each one matters.
Score a page on whether an assistant could actually answer with it, across four pillars worth a hundred points.
Find out what is stopping a page being indexed, which is the half of the question anything can actually answer.
See the answer an engine can build from your page, with every sentence labelled by the element it came from.