Citebase
Guides Pay $197

OpenAI’s own words

How to show up in ChatGPT search

OpenAI says any public website can appear in ChatGPT search. The one requirement it states is that you do not block its search crawler, OAI-SearchBot. It does not publish ranking factors, and it says placement is not guaranteed. Everything past the crawler is ordinary work on pages a buyer would trust.

What OpenAI documents

Three OpenAI pages cover this. Here is what each one says, without additions.

Publishers and Developers FAQ
Any public website can appear in ChatGPT search. To be included in summaries and snippets, do not block OAI-SearchBot. You may need to update robots.txt. ChatGPT adds utm_source=chatgpt.com to referral URLs, so you can see the traffic in analytics. To keep content out of training, disallow GPTBot, which is a separate setting.
Overview of OpenAI Crawlers
OAI-SearchBot supports ChatGPT search. Sites that opt out of it will not be shown in ChatGPT search answers, though they can still appear as a navigational link. A robots.txt change can take about 24 hours to take effect for search. OpenAI publishes the crawler’s IP ranges at openai.com/searchbot.json.
Searching the web with ChatGPT
ChatGPT may search automatically when a question would benefit from current information, or a person can start a search from the tools menu. It usually rewrites the question into one or more queries, and sometimes sends them to other search providers. Ranking uses multiple factors meant to help people find relevant, reliable information. Placement is not guaranteed. Responses that use search may include citations, and a Sources view lists cited and related links. OpenAI also says results and citations can be incomplete, outdated, or incorrect.

That is the whole official picture. None of these pages lists the ranking factors, offers a submission form for ordinary sites, or describes a way to pay for placement in search answers.

Step 1: let OAI-SearchBot in

Open https://yourdomain/robots.txt in a browser and read what is actually there. If there is no group for OAI-SearchBot, the User-agent: * group applies to it. If that group disallows the paths you care about, ChatGPT search cannot use them.

A group of its own removes the guesswork:

User-agent: OAI-SearchBot
        Allow: /

Then check the layer in front of the site. A bot-protection setting at your host or CDN can block a crawler that robots.txt allows. OpenAI publishes the IP ranges for OAI-SearchBot so you can write an allow rule that checks both the user agent and the address. Look at your server or CDN logs for OAI-SearchBot requests and their status codes. A wall of 403 responses is your answer.

GPTBot is a different question. It is OpenAI’s training crawler. OpenAI says the two settings are independent, so you can allow OAI-SearchBot and disallow GPTBot. The crawler guide has both on one table.

ChatGPT-User is a third agent, for visits a person triggers, such as asking ChatGPT to open a link. OpenAI says it is not automatic crawling and that robots.txt rules may not apply to it. It is not how search inclusion is decided.

noindex, and pages you want kept out

The publisher FAQ has one more detail worth knowing. If you disallow OAI-SearchBot but a page is still found through a third-party search provider or other crawling, ChatGPT Atlas, OpenAI’s browser, may show only the link and page title. To keep a page out entirely, OpenAI says to use a noindex meta tag, and that the crawler has to be allowed to read the page to see that tag. So a robots.txt disallow and a noindex tag do different jobs. Use the disallow to keep content out of summaries and snippets. Use noindex, with crawling allowed, to keep the page from being listed at all.

Step 2: be findable by the search behind it

OpenAI says ChatGPT search sometimes uses other search providers, and its help article points to Microsoft’s and Shopify’s privacy policies for how those providers handle queries. It does not say which provider answers which query, or how results from its own crawler and from partners are combined.

So the following is a precaution, not a documented rule: make sure Bing can index your important pages. Verify the site in Bing Webmaster Tools, submit a sitemap, and consider IndexNow, a protocol for telling participating search engines when a URL is added, changed, or removed. IndexNow’s documentation is clear that a successful submission only means the URL was received, not that it will be indexed.

Being indexed in Google matters for Google’s own AI features. It is not something OpenAI names as a factor for ChatGPT.

Step 3: write pages that answer the question

ChatGPT turns a buyer’s question into search queries. The page that comes back has to answer something close to it. This part is observed practice, not OpenAI guidance:

  • Say what the product is in the words a buyer uses. “Scheduling page for consultants” is matchable. “Time, reimagined” is not.
  • Put who it is for and what it costs in text, near the top, on a page that is linked from the homepage.
  • Keep the pricing page current. OpenAI warns that answers can be outdated. Old prices on your own site make that more likely.
  • If you publish a comparison page, make it fair and dated. A page that only praises you is less useful to anyone, including a system that is trying to summarize the category.

Pages written about you by others also matter. If ChatGPT cites a roundup for your category and you are missing from it, the competitor guide covers what to check.

Step 4: if you sell physical products

Shopping results work a little differently. OpenAI’s shopping help article says product results are selected independently by ChatGPT and are not ads. It says ChatGPT weighs structured metadata from first- and third-party providers, and that merchants are ranked on factors like availability, price, and quality. Shopify merchants are already integrated through Shopify Catalog. Other merchants can apply to send OpenAI a product feed. Product data and feeds are a topic of their own, and they get a separate guide later in this series.

Step 5: measure what arrives

Because ChatGPT adds utm_source=chatgpt.com to referral links, visits from cited links show up with that source in most analytics tools. In Google Analytics 4, the default channel group includes an AI Assistant channel, and Google names ChatGPT among the sources it covers.

Two limits to keep in mind. First, a click is not the same as being mentioned. An answer can recommend you by name with no link, and nobody clicks. Second, people who copy your name out of an answer and search for it later arrive as a branded search or a direct visit. Analytics cannot tie that back to ChatGPT. Reading the answers is the only way to see the mentions.

What will not work

  • Paying for a better position in search answers. OpenAI says shopping results are not ads and are not influenced by its partnerships. ChatGPT ads exist, and OpenAI has a separate crawler, OAI-AdsBot, that only visits landing pages submitted as ads.
  • An llms.txt file as an entry ticket. OpenAI’s crawler documentation does not mention llms.txt. Publishing one is harmless. It is not the opt-in.
  • Asking ChatGPT to remember you. Telling your own ChatGPT account about your product changes your own answers. It does not change what a stranger sees.
  • Treating one answer as a ranking. OpenAI says placement is not guaranteed, and answers change between runs. Run the same question several times, logged out, before you conclude anything. The manual check has a method.

A short checklist

  1. robots.txt allows OAI-SearchBot, in a group of its own.
  2. Your firewall or CDN lets OAI-SearchBot through. Logs show 200 responses.
  3. The site is verified in Bing Webmaster Tools with a current sitemap.
  4. The homepage and pricing page say what the product is, who it is for, and what it costs, in text.
  5. Analytics shows a chatgpt.com source, or the AI Assistant channel in GA4.
  6. You have a saved set of buyer questions and the answers from today, so next month has something to compare against.

Related: what AEO means, and AI search vs SEO.

Sources