Home / GEO and AI search visibility / ChatGPT search
how ChatGPT cites websites
ChatGPT search
ChatGPT search is the part of ChatGPT that looks up current information on the web and answers with links to the sources it used.
For a B2B site owner the useful questions are narrow: how does it find pages, which crawler setting controls inclusion, how do you see the visits it sends and what does OpenAI simply not say? This page answers those from OpenAI's own documentation, and sits under our GEO agency work on visibility across AI assistants.
Configure crawler access deliberately, measure ChatGPT referrals in your analytics and set realistic expectations about what can and cannot be influenced.
ChatGPT search leverages third-party search providers and content provided directly by partners; the launch search model was a fine-tuned GPT-4o. OpenAI: Introducing ChatGPT search ↗
OpenAI documents OAI-SearchBot for ChatGPT search and GPTBot for training, and states each robots.txt setting is independent. OpenAI: overview of OpenAI crawlers ↗
New to the topic? Read the background on artificial intelligence.
how ChatGPT cites websites
how does ChatGPT search find sources?
Through third-party search providers, content from partners and OpenAI's own search crawler.
OpenAI's launch post, Introducing ChatGPT search, says the feature leverages third-party search providers as well as content provided directly by its partners, and that the search model was a fine-tuned version of GPT-4o at launch.
OpenAI's crawler overview adds its own crawler: OAI-SearchBot is used to surface websites in search results in ChatGPT's search features.
That combination matters. A page that ranks in a conventional search index and is reachable by OAI-SearchBot has two routes into a ChatGPT search answer; a page invisible to both has neither.
OpenAI does not publish how it ranks or chooses among candidate pages, or which third-party providers are used for which queries. Treat claims of a precise 'ChatGPT search ranking factor' as opinion unless they cite OpenAI.
Third-party search providers.
Partner content.
OAI-SearchBot crawling.
how ChatGPT cites websites
how do we make sure our site can appear in ChatGPT search?
Do not block OAI-SearchBot, and allow requests from OpenAI's published IP ranges.
OpenAI's publishers and developers FAQ says any public website can appear in ChatGPT search, and that for content to be included in summaries and snippets you must not block OAI-SearchBot, updating robots.txt if necessary.
The crawler overview recommends allowing OAI-SearchBot in robots.txt and allowing requests from OpenAI's published IP ranges. Many B2B sites sit behind a CDN or firewall with bot rules that block unfamiliar agents by default, so check there as well as robots.txt.
OpenAI says it can take about 24 hours from a robots.txt update for its search systems to adjust, so test a day after any change rather than immediately.
Beyond access, the content advice is the same as for any retrieval-based assistant: a direct answer near the top, specific facts with sources and stable details about who you are. Those page-level changes are covered in LLM SEO and answer engine optimisation.
how ChatGPT cites websites
can we appear in ChatGPT search but stay out of training?
Yes. OpenAI states the search and training settings are independent.
OpenAI documents GPTBot as the crawler for training its generative AI foundation models and OAI-SearchBot for search, and says each robots.txt setting is independent: a site can allow OAI-SearchBot to appear in search while disallowing GPTBot.
It also says that if both are allowed it may use one crawl for both purposes to avoid duplicate crawling. That is an efficiency note, not a change to what each setting permits.
A third agent, ChatGPT-User, visits pages when a user's request triggers it. OpenAI says it is not used for automatic crawling, that robots.txt rules may not apply to it because the actions are user-initiated, and that it is not used to decide whether content appears in search.
Many firms want the opposite of a blanket block: present in answers, out of training. Write that decision down per crawler, as recommended in our GEO audit, so the next website change does not undo it.
| Check | What to establish |
|---|---|
| 1 | GPTBot is for training. |
| 2 | OAI-SearchBot is for search. |
| 3 | ChatGPT-User acts on user requests. |
how ChatGPT cites websites
how do we opt out of ChatGPT search?
Disallow OAI-SearchBot, and use noindex if you also want links and titles suppressed.
OpenAI's crawler overview says sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though they can still appear as navigational links.
The publishers FAQ adds that if OpenAI obtains the URL of a disallowed page from a third-party search provider or by crawling other pages, and has signals it is relevant, it may surface just the link and page title. To prevent that, OpenAI says to use the noindex meta tag, the same directive defined in Google's robots meta tag specification.
There is a catch the FAQ states plainly: for the crawler to read a meta tag, it must be allowed to crawl the page. Blocking in robots.txt and relying on noindex at the same time does not work, because the tag is never seen.
Perplexity documents a similar split between PerplexityBot and Perplexity-User, so whatever policy you choose for ChatGPT search is worth applying consistently across assistants.
how ChatGPT cites websites
how do we track traffic from ChatGPT search?
Look for utm_source=chatgpt.com on landing URLs in your analytics.
OpenAI's publishers FAQ says ChatGPT automatically includes the UTM parameter utm_source=chatgpt.com in referral URLs, so publishers can track inbound traffic from ChatGPT search results in analytics platforms such as Google Analytics.
In Google Analytics that usually appears as a session source of chatgpt.com. Build a saved exploration or a custom channel group for it, and add other assistants' referrers alongside so the trend is read together.
What you will not get is anything like Search Console, where Google reports AI feature traffic inside the Web search type: no impressions, no queries, no position. You see visits that arrived, not the answers in which you were mentioned without a click.
That gap is why a GEO baseline also checks a fixed list of buyer questions by hand on a schedule. Referral data shows the clicks; the prompt baseline shows the mentions and how you were described.
Filter on utm_source=chatgpt.com.
No impressions or query data.
Pair referrals with a prompt baseline.
how ChatGPT cites websites
what is still unknown about ChatGPT search?
How sources are ranked, how often answers change and how many people saw an answer naming you.
OpenAI documents access, opt-out and the UTM parameter. It does not document a ranking method, a reporting tool for site owners or a way to request inclusion.
Answers also vary by phrasing, account context and date, so a single screenshot proves little. A useful test fixes the questions and the date and repeats them.
Be cautious of anyone promising a ChatGPT search placement. The honest offer is to remove access barriers, improve the pages most likely to be retrieved and measure the result, which is how we run AI SEO agency work.
If you would like a dated baseline of how ChatGPT search describes your firm against the questions your buyers ask, contact Predeploy AI.
No published ranking method.
Answers vary run to run.
No guaranteed placements exist.
Make the decision concrete
What does the current task cost in time?
Choose a task connected to chatgpt search and enter its current workload. This gives you a baseline to discuss before buying training or commissioning a build.
A baseline is useful even if you decide to keep the process manual. Measure the work that matters, then compare quality and review effort as well as time.
Assess the workflow before changing it, or plan business AI training if the gap is how people use existing tools.
Your current workload
Estimate how much time the existing task takes before discussing chatgpt search.
16 hours per week
People × minutes × weekly tasks ÷ 60. A workload baseline from your inputs, not a prediction of AI time savings. Include review and support time when comparing a proposed workflow.
A useful AI investment leaves your team able to do the work and judge the result.
Continue with the right question
Where should you go next?
Use the overview for context, then follow the pages that match your delivery decision.
See our ai powered gtm service or read the geo and ai search visibility overview.
FAQ
Questions, answered
Evidence and further reading
Which sources support the guidance?
Source links support the specific background described here. The practical checklists and illustrative examples are predeploy’s guidance.
ChatGPT search leverages third-party search providers and content provided directly by partners; the launch search model was a fine-tuned GPT-4o. OpenAI: Introducing ChatGPT search ↗
OpenAI documents OAI-SearchBot for ChatGPT search and GPTBot for training, and states each robots.txt setting is independent. OpenAI: overview of OpenAI crawlers ↗
Any public site can appear in ChatGPT search if OAI-SearchBot is not blocked; disallowed pages may still surface as link and title unless noindex is used (which needs crawl access); ChatGPT adds utm_source=chatgpt.com to referral URLs. OpenAI Help Center: publishers and developers FAQ ↗
Perplexity documents PerplexityBot for surfacing and linking sites in search results, separate from user-triggered Perplexity-User fetches. Perplexity: Perplexity crawlers ↗
Background reading on the broader field of artificial intelligence. Wikipedia: artificial intelligence ↗
Choose your next step
Explore geo and ai search visibility
Start with a topic overview, or go directly to the question your team needs to answer.
GEO and AI search visibility
Be found, described and cited in AI answers as well as search.
Start with the decision
More questions in this topic
