How to Check If ChatGPT Cites Your Website

To check whether ChatGPT cites your website, the fastest route is to run a free audit — it queries the engines for you and tells you whether your domain came back as a cited source, with no login and no setup. The manual route is to open ChatGPT with web search enabled, ask the questions your customers ask, and expand the source list under each answer to see whether any link points at your domain. Both are worth doing, and this guide covers each in turn.

A note on scope, so you can jump to the right page. This article is narrow on purpose: how to check one engine — ChatGPT — for one site, yours. If you want the wider view, three companion pieces cover it: how to check AI search visibility covers the same job across all engines, how to get cited by ChatGPT covers earning citations rather than measuring them, and citation tracking for Perplexity and ChatGPT plus the AI citation tracking guide cover doing this continuously instead of once.

The fastest method: run a free audit

Enter your domain and run the audit. It reports whether your site came back as a cited source across four engines — ChatGPT, Perplexity, Gemini and Claude — as a single line: cited by N of 4. There is no account, no card, and no crawl to wait through.

Check your domain now → — or deep-link straight to a result with your domain prefilled, for example /free-audit?url=yourdomain.com.

Two things about the report are worth understanding before you read yours.

It is one run, not a status. The audit is explicit that it is a one-time snapshot, and that honesty matters here more than it does in classic SEO. Ahrefs measured 45.5% of citations replaced over its November 2025 window, and only 13.7% source overlap when the same question was asked repeatedly across 540,000 queries (Ahrefs, via our statistics roundup). SparkToro found in January 2026 that asking an identical question 100 times often surfaced a given source fewer than 1 in 100 times. So a "not cited" result is real information, but it is one draw from a shuffled deck — the reliable read comes from repeating the check on a schedule.

Four engines is deliberate. Checking ChatGPT alone gives you a partial picture, because 91% of cited sources appear in only one engine (Growth Memo, May 2026 — see the data). A site invisible in ChatGPT is quite often cited in Perplexity, and the reverse. Reading the four side by side is what turns "we're not in AI answers" into "we're absent from these engines specifically."

If you are not cited, the report doesn't stop at the bad news — it walks through the three stages a page has to clear before any engine can cite it. That framework is the last section of this article.

What "cited" actually means

A citation is not a mention. ChatGPT can name your brand in a sentence without linking to you at all, and it can link to a page on your domain without ever naming you. These are different outcomes with different fixes, and conflating them is the most common measurement error.

Three things are worth separating every time you check:

The distinction also explains a result that surprises people: your competitor's review page can be the cited source for a question about your product. That is normal. Roughly 85% of cited pages are third-party rather than brand-owned (eMarketer, January 2026 — source), which is why "we're not cited" often means "someone else is being cited about us" rather than "we're invisible."

The manual method: check it yourself

The manual check is slower and noisier than the audit, but it shows you something the audit can't: the actual wording of the answer you're losing to. Do it once for your five most commercially important questions.

1. Turn on web search. In ChatGPT, make sure the answer is drawing on live web results rather than model memory. Without retrieval there are no sources to inspect, and an answer that names your brand from training data is not a citation.

2. Ask the questions your buyers ask, not your brand name. "What is [your brand]" almost always surfaces your homepage and tells you nothing. The queries that matter are the ones with commercial intent and no brand attached: "best tools for [your category]", "how do I [the problem you solve]", "[competitor] alternatives". Those are the answers where a citation is genuinely contested, and they are the ones brands most often lose: our AI visibility benchmark measures how many brands an engine will name outright still go missing when that same engine is asked for alternatives in their category.

3. Expand the sources. Open the source list under the answer and read the domains. You are looking for one thing: is any listed URL on your domain? Note the exact URL, not just the domain — being cited for a blog post and being cited for your pricing page are very different outcomes.

4. Check the URL for a tracking parameter. Links surfaced by ChatGPT frequently carry a ?utm_source=openai parameter. This is genuinely useful: it means you can segment those visits in your analytics later, and it is the cleanest way to confirm a click came from an AI answer rather than a search result.

5. Ask again tomorrow. Same question, same wording. If the source list changed — and per the Ahrefs numbers above it very often does — you have learned that your one-off check was measuring noise. That realization is usually what moves teams from spot-checking to scheduled tracking.

One more manual check is worth running once: confirm the crawlers are even allowed in. If your robots.txt blocks GPTBot or OAI-SearchBot, no amount of content work will produce a citation, because the page was never fetchable. The free audit checks this automatically, but it takes thirty seconds to verify by hand. Which crawlers are worth allowing is an empirical question rather than a list to guess at — our AI crawler traffic stats record which bots actually reach a live site, and in what proportion.

Cross-check it in Google Search Console

Here is the honest limitation first, because it is the part most guides skip: Search Console cannot show you ChatGPT citations. It reports Google Search, full stop. Anyone telling you to find ChatGPT data in GSC is describing something that does not exist.

What Search Console can give you is an adjacent signal on the Google side, and it is a useful one because the underlying cause usually overlaps.

Look for queries with high impressions and near-zero clicks. You are ranking, users are seeing you, and nobody is clicking. In 2026 that pattern often means an AI Overview answered the question directly above your result. The supporting data is stark: 93% of AI Overview impressions produce no click (Semrush, September 2025), and Ahrefs measured a 58% CTR drop to the #1 organic result when an AI Overview appears (both in our statistics roundup).

If you have connected Search Console to TriRank, this is a built-in view rather than a spreadsheet exercise — the High impressions, low CTR tab on the Keywords page surfaces exactly these queries. The product is deliberately hedged about what the pattern proves, and so is this article: high impressions with low CTR often signals an AI Overview is capturing the intent, but it can equally be a weak title or meta description. We do not inspect the SERP for an AI Overview, so this is an interpretation of the impressions-and-clicks shape, never a detection claim.

Why bother, if it's a different engine? Because the content weakness is usually shared. A page that a Google AI Overview summarizes without crediting is typically the same page ChatGPT declines to cite: the answer is buried, the claim is unsourced, and there is nothing quotable near the top. The GSC view gives you a prioritized list of which questions to fix first, ranked by how much attention they already command.

What to do if ChatGPT isn't citing you

A page missing from an AI answer almost always failed at one of three stages, and they run in order — each one only matters once the previous one passes. This is the same framework the free audit uses when it reports a "not cited" result.

Stage 1 — Access. Can the crawler reach the page? An engine can only cite a page it is allowed to fetch. A robots.txt rule against GPTBot, ClaudeBot, PerplexityBot or Google-Extended, a noindex tag, or content sitting behind a login removes the page from consideration before anything else is weighed. This is the cheapest stage to fix and the most embarrassing one to fail.

Stage 2 — Extraction. Can the content be parsed once fetched? A fetched page still has to yield clean, machine-readable text. Copy that only appears after client-side rendering, an unclear heading structure, or missing structured data all make the answer harder to lift cleanly. Adding statistics, quotes and structured data lifted citation likelihood by about 40% in joint Princeton and Georgia Tech research (see the data).

Stage 3 — Grounding. Is the page worth citing for that question? Reachable and parseable is not the same as chosen. Engines tend to cite the source that answers the question most directly and carries the clearest authority signals: a specific answer near the top of the page, a stated author or methodology, and corroboration elsewhere on the web. Placement is measurable here — 44.2% of cited content sits in the first 30% of a page (Growth Memo, February 2026), and 52.2% of cited passages contain original data (Search Engine Land).

Stage 3 is where most sites actually lose, and it is the one a single scan can see the least of. It is also the stage that closes the loop: once you know which questions you're absent from, the fix is writing the page that answers them in a form an engine can lift — a direct answer up top, a real source under every number, and a structure that survives extraction. TriRank's engine is built to run that loop, but the diagnosis is the part you should do first, and it's free.

Two closing cautions, in the spirit of measuring honestly. First, nothing on this list can promise a citation — which sources an engine picks varies by question, by engine, and between runs of the same query. Second, do not judge the work by one check. With 45.5% of citations replaced over a measurement window and only 14% of brands tracking their citations at all (Goodfirms, 2026), the durable advantage is not a single good result — it's being one of the few teams that measures repeatedly and can tell a real gain from churn.

Run your free audit to get the baseline, then decide whether it's worth watching every week.

FAQ

How do I check if ChatGPT cites my website? The fastest way is to run a free audit that queries the engines for you and reports whether your domain came back as a cited source — no login required. The manual alternative is to open ChatGPT with web search enabled, ask the questions your buyers ask, and expand the source list under each answer to see whether any link points at your domain. Do both: the audit gives you a baseline in seconds, and the manual pass shows you the wording of the answer you're competing with.

Is my website cited by AI? Most sites are not. A Search Engine Journal study of 177 brands (May 2026) found roughly 90% have zero presence in AI-generated answers. The only way to know for your own domain is to check it, because citation is decided per question rather than per site — you can be cited for one query and absent from the next.

Why does ChatGPT cite my site one day and not the next? Because citation sets are unstable by design. Ahrefs measured 45.5% of citations replaced over its November 2025 window and only 13.7% source overlap when the same question was asked across 540,000 queries. A single check is a snapshot, not a status — a repeated, scheduled check is the only way to tell a real trend from normal churn.

Can Google Search Console show me ChatGPT citations? No. Search Console reports Google Search only, so ChatGPT citations never appear there. What it can show is an adjacent signal on the Google side: queries with high impressions and near-zero clicks, which often means an AI Overview answered the question before anyone clicked. That is a different engine, but usually the same underlying content weakness.

Does ChatGPT send traffic when it cites you? Sometimes. When a user clicks a cited link, the visit shows up in your analytics as a referral from chatgpt.com, and the destination URL frequently carries a utm_source=openai parameter — which makes those visits easy to isolate. But a citation with no click still counts as visibility, so referral traffic undercounts how often you are actually being cited.

Related reading

• How to Check AI Search Visibility — the same job, across every engine
• How to Get Cited by ChatGPT: An 8-Step Playbook — earning citations, not measuring them
• Citation Tracking for Perplexity and ChatGPT — doing this continuously
• The AI Citation Tracking Guide — the tracking discipline in full
• AI SEO Statistics — every figure cited above, with sources