Google & AI Search · 5 min read
How Does ChatGPT Search Pick and Cite Sources?
Short answer
ChatGPT search draws on pages that OpenAI's OAI-SearchBot crawler can reach, along with outside search results (Microsoft announced Bing as its default search experience in 2023), then writes an answer with links to some of them. To be eligible, allow OAI-SearchBot in robots.txt, make sure Bing indexes your pages, and write passages that answer questions directly.
Key takeaways
- Blocking OAI-SearchBot removes your pages from ChatGPT search answers; blocking GPTBot only opts you out of model training.
- Bing indexing is a sensible precaution because Microsoft announced Bing as ChatGPT's default search experience in 2023, though OpenAI does not publish its full provider list.
- You can make a page eligible and easy to quote, but you cannot control whether a given answer cites it.
A prospect asks ChatGPT which vendors to shortlist, and the answer links three sites. If yours isn't one of them, the first question is whether ChatGPT could have used your pages at all. This page covers the crawlers involved, the robots.txt rules that control them, where search providers fit, and the steps that make a site eligible.
Which OpenAI crawlers visit your site?
As of November 2026, OpenAI documents four user agents, and OAI-SearchBot is the one that decides whether your pages can appear in ChatGPT search.
Each has a separate job and a separate robots.txt rule:
| User agent | What it does | robots.txt effect |
|---|---|---|
| OAI-SearchBot | Finds pages to show in ChatGPT search | Disallow it and your pages are not shown in search answers |
| ChatGPT-User | Visits a page when a user's request calls for it | Rules may not apply, because a person triggered the visit |
| GPTBot | Collects content that may train foundation models | Disallow it to opt out of training |
| OAI-AdsBot | Checks the safety of pages submitted as ChatGPT ads | Only relevant if you advertise |
OpenAI says these settings are independent. You can block GPTBot for training and still allow OAI-SearchBot for search. One detail: a site that blocks OAI-SearchBot can still appear as a plain navigational link, just not as a source in answers.
How do you control access with robots.txt?
Add explicit rules for each OpenAI user agent, and expect changes to take about a day to apply.
A typical file for a business that wants search visibility looks like this:
User-agent: OAI-SearchBot
Allow: /
User-agent: GPTBot
Allow: /
Change the GPTBot line to Disallow: / if you don't want your content used for training. OpenAI says it can take roughly 24 hours after a robots.txt update for its systems to adjust. Also check your CDN and firewall: bot-protection features sometimes block these crawlers even when robots.txt allows them. Our GEO readiness checker reads your robots.txt and reports which AI crawlers are allowed.
Does ChatGPT search use Bing?
Bing very likely plays a part, but OpenAI does not publish a full list of the search providers behind ChatGPT search, so treat Bing indexing as a precaution rather than a guarantee.
In May 2023 Microsoft announced that Bing was coming to ChatGPT as its default search experience. OpenAI now runs its own search crawler as well, and its current crawler documentation does not name outside providers. What we can say with confidence: pages that are well indexed in Bing and reachable by OAI-SearchBot have two routes to being found, while pages missing from both have none.
How do ChatGPT citations appear?
Citations appear as links attached to parts of the answer, and the set of sources can change each time the same question is asked.
The exact layout changes as OpenAI updates the product, so describe it to stakeholders loosely. Answers generally show small source links next to the claims they support, plus a list of the sources used. Clicks reach your site as referrals from chatgpt.com, and outbound links are often tagged with utm_source=chatgpt.com. Confirm that pattern in your own landing-page data before you build reports on it; our guide to tracking AI search traffic in GA4 shows how to set up the channel.
What can you do to become eligible?
Make the page reachable, indexed and easy to quote, then let the answers decide.
- Allow OAI-SearchBot in robots.txt and at the CDN level.
- Verify the site in Bing Webmaster Tools, submit a sitemap and fix any crawl errors it reports.
- Turn on IndexNow so new and changed URLs are pinged to Bing and other participating engines (Naver, Seznam.cz, Yandex and Yep support it).
- Put the direct answer to each question in the first sentence under its heading.
- Keep facts specific and dated, with a named source, so the passage stands alone when quoted.
- Make sure key text is in the HTML rather than behind tabs that load only on click.
These steps overlap heavily with normal SEO. For the wider picture, read what generative engine optimization is.
What can't you control?
You cannot control which sources a particular answer cites, how it phrases your information, or whether the user clicks.
ChatGPT tailors answers to the wording of the prompt and the conversation so far. Two people asking nearly the same question can see different sources. Answers can also summarize you without sending a visit.
That means you should judge results over many prompts and several weeks. Sample a fixed set of buyer questions each month, record which domains are cited, and compare that with referral sessions and leads. If you'd like that tracking set up alongside the technical fixes, see our GEO and AI search service.
Frequently asked questions
Does blocking GPTBot keep my site out of ChatGPT search?
No. OpenAI says GPTBot and OAI-SearchBot are controlled separately. GPTBot covers training data for foundation models, while OAI-SearchBot decides whether your pages can be shown in ChatGPT search answers.
How long does a robots.txt change take to affect ChatGPT search?
OpenAI says it can take about 24 hours after a robots.txt update for its systems to adjust to OAI-SearchBot rules. Plan changes with that delay in mind.
Why does ChatGPT-User visit pages I blocked?
ChatGPT-User fetches a page when a person in ChatGPT asks it to, and OpenAI notes that robots.txt rules may not apply to these user-initiated visits. It is not used to decide what appears in search.
Can I pay to be cited in ChatGPT search?
There is no paid route into the organic answer. OpenAI separately runs an ad-safety crawler, OAI-AdsBot, for ads shown in ChatGPT, which is a different system from search citations.