ChatGPT Citations: When ChatGPT Links a Source, and Whether It Can Link Yours
ChatGPT citations are the source links ChatGPT attaches to an answer, and they appear only when ChatGPT searches the web. An answer from memory has no web sources behind it. If it lists references anyway, OpenAI's own help center warns they can be invented.
That leaves two practical questions. If you are reading an answer: when is a ChatGPT citation real? If you run a website: what decides whether ChatGPT can cite yours? OpenAI documents more of this than most people think, and none of it is a ranking trick.
When does ChatGPT cite sources?
When it searches. OpenAI's search help page says "ChatGPT may search the web automatically when your question would benefit from current information," and you can also pick Search from the tools menu. Then: "Responses that use web search may include citations. Select a citation to open its source."
Without search, there is nothing to link to. From OpenAI's accuracy FAQ: "Without search enabled: Responses are based on what the model has learned during training." And the models "are trained on data up to a certain point and responses do not incorporate information about events beyond that, unless tools are used."
| Answer type | Cites sources? | What shapes it |
|---|---|---|
| From memory | No web sources | Training data, up to the model's knowledge cutoff |
| Web search | Yes: responses "may include citations" | The pages ChatGPT's search finds and can read |
| Deep research | Yes: "multi-source, cited answers" | Many sources, across the web and internal knowledge |
An answer from memory cannot cite you. An answer from search can only cite pages ChatGPT was able to fetch.
Does ChatGPT make up citations?
It can. OpenAI lists "Fabricated quotes, studies, citations or references to non-existent sources" among the hallucinations ChatGPT produces. Search does not make every link right either: "Search results and citations can be incomplete, outdated, or incorrect." OpenAI's advice is to open the cited source and check that it supports the answer. A reference you cannot click through to is the one to check first.
How ChatGPT search decides what to cite
OpenAI discloses three things about the process:
- It rewrites the question. "ChatGPT search typically rewrites your query into one or more targeted queries that it sends those providers." OpenAI's example turns a long question about cancer drugs into "CCR8 immunotherapy drug development 2025," and may then send narrower follow-up queries after reading the first results.
- It uses search partners as well as its own crawler. "ChatGPT search sometimes partners with other search providers." Its own crawler is OAI-SearchBot.
- It ranks, without saying how. "ChatGPT ranks search results using multiple factors intended to help users find relevant, reliable information. Placement is not guaranteed."
That is the whole public ranking disclosure. The prompt your customer types is not the query ChatGPT searches, so anyone promising the exact phrase that earns a ChatGPT citation is guessing.
What decides whether ChatGPT can cite your site
Here OpenAI is specific. Its eligibility rule: "To make a website eligible for inclusion, allow OAI-Searchbot to crawl the site and confirm that the website host or content delivery network allows traffic from OpenAI's published searchbot IP addresses." Two conditions, joined by "and". Add a third that comes from how the crawlers work:
- Robots.txt allows OAI-SearchBot. From OpenAI's crawler documentation: "Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though can still appear as navigational links." Changes are not instant: "it can take ~24 hours from a site's robots.txt update for our systems to adjust."
- Your firewall or CDN lets OpenAI through. Robots.txt can say yes while a bot-protection rule answers with a 403 or a challenge page. Your own browser visit still works, so the block is easy to miss.
- The facts are in the HTML. Vercel's December 2024 crawler study found that "none of the major AI crawlers currently render JavaScript," and named OAI-SearchBot, ChatGPT-User and GPTBot among them. A price that appears only after a script runs may not exist for them.
OpenAI lists the same failure modes from the other side. Its accuracy FAQ says the model "may not be able to obtain relevant information from a given website due to technical issues, paywalls or preferences set via robots.txt."
OpenAI's crawlers, and what blocking each one does
| Crawler | What it does | If you block it |
|---|---|---|
| OAI-SearchBot | Surfaces sites in ChatGPT search | Your pages "will not be shown in ChatGPT search answers," though can still appear as navigational links |
| GPTBot | Crawls content that may be used to train OpenAI's models | Signals your content should not be used for training. Search is a separate setting |
| ChatGPT-User | Visits a page when a user's request needs it | "robots.txt rules may not apply." Not used to decide what appears in search |
| OAI-AdsBot | Checks landing pages submitted as ads | Matters only if you advertise: it visits pages submitted as ads, and nothing else |
GPTBot and OAI-SearchBot are easy to confuse, and they do different jobs. OpenAI: "Each setting is independent of the others." Blocking GPTBot does not remove you from ChatGPT search, and OpenAI describes GPTBot as a training crawler, not a route into citations. More in GPTBot vs OAI-SearchBot and our AI crawler list.
What you control is reachability, not ranking. Passing these checks makes your pages eligible. OpenAI says placement is not guaranteed, and no markup, file or phrasing has published evidence that it earns a citation. See how LLMs prioritise information about brands for the evidence.
How to see ChatGPT citations in your analytics
OpenAI's publisher FAQ stated that "ChatGPT automatically includes the UTM parameter utm_source=chatgpt.com in referral URLs" (archived copy, 4 October 2026). That FAQ sat in OpenAI's section for its Atlas browser, and its address now redirects to the Atlas retirement notice. Treat the tag as behaviour to confirm in your own data, not a guarantee.
Filter your analytics for that source and count it as a floor. A reader who sees your name in an answer and types your address arrives as direct traffic. Clicks on ChatGPT ads are separate: they carry whatever tracking parameters you add to the ad's landing URL, so paid and cited traffic stay apart.
A 15-minute ChatGPT citation check
- Read your robots.txt. Confirm no rule disallows OAI-SearchBot, including a blanket rule under "User-agent: *".
- Ask about your firewall. Whoever runs your CDN or bot protection should confirm that OpenAI's published searchbot IP ranges are allowed, not challenged.
- View the source of your key pages. Prices, product names and company facts should be in the raw HTML, not loaded later by a script.
- Ask ChatGPT five buyer questions with Search on. Note who gets cited, then open each citation and check it says what the answer claims.
- Check your analytics for utm_source=chatgpt.com, and compare month to month.
Other AI platforms follow their own rules; for one of them, see how Perplexity cites sources.
Doing this for one site on one platform is manageable. 99Visibility covers five AI platforms (ChatGPT, Gemini, Perplexity, Google AI Overviews and Claude): it checks whether AI crawlers can reach your site, checks the answers against the brand facts you enter, and tracks your competitors. Your first audit is free, with no credit card, and covers ChatGPT and Google AI Overviews. See the plans for all five platforms. Inside the dashboard, Agent V reads your audits and your site and drafts the fix, such as page copy for you to paste. It never edits your site itself.
FAQ
Does ChatGPT cite its sources?
When it searches the web, yes: responses that use search "may include citations." When it answers from training data alone, it has no web sources to cite, and any references it lists need checking.
Why does ChatGPT cite my competitor and not me?
OpenAI says it ranks results "using multiple factors" and that placement is not guaranteed. Rule out the causes you control first: OAI-SearchBot blocked in robots.txt, a firewall blocking OpenAI's IP ranges, or facts that only appear after JavaScript runs.
Can I pay ChatGPT to cite my website?
No. ChatGPT ads appear below the answer, labelled as sponsored, and OpenAI states that "Ads do not influence ChatGPT's answers."
Find out whether ChatGPT can cite you. Run a free AI visibility audit: no credit card needed.
Want more of our research in Google Search and AI Mode? Add us as a preferred source.
See what AI says about your brand
Your free audit asks ChatGPT and Google AI Overviews about your brand. Paid plans add Gemini, Perplexity and Claude, and show every answer that contradicts your facts.
Start Free Audit- ChatGPTMentioned, 2nd of 5 brands
- Google AI OverviewsNot mentioned
- PerplexityMentioned and cited as source [1]
- ClaudeSays $29/mo. Your pricing page says $49/mo.
Related Articles
ChatGPT Ads: How They Work, What They Cost, and Why an Ad Is Not the Answer
ChatGPT ads run below the answer for Free and Go users, bought by auction in OpenAI Ads Manager from $25 a day. How they work, what they cost, and why paying never puts you inside the answer.
GPTBot vs OAI-SearchBot vs ChatGPT-User: Which OpenAI Crawler to Allow
GPTBot collects training data. OAI-SearchBot decides whether ChatGPT search can show your site. ChatGPT-User fetches pages live, and robots.txt may not apply to it. What OpenAI says each one does, a decision table by goal, and the firewall rule robots.txt cannot fix.
Perplexity SEO: How Perplexity Finds and Cites Sources
Perplexity searches the web for its answers and cites sources by number. What its own docs say decides access, which crawler ignores robots.txt, and what to check.