How to get your website on ChatGPT

You can’t buy the mention.You can earn the citation.

ChatGPT’s answers with links come from its search feature, which reads pages found through search providers and OpenAI’s own crawler, OAI-SearchBot. You can’t pay for a place in the answer or force one. What you can do: let the right crawler in, describe your business plainly enough to be quoted, and be mentioned by the sites a search turns up. Then measure the visits, because ChatGPT tags its links with utm_source=chatgpt.com.

Check my readiness Start free
OpenAI crawler names Checked 29 October 2026 no invented stats

By the Builderify team

Made-up business. Quoted because the page said it: the service, the area, the price.
01 / How it works

How does ChatGPT decide
which websites to link?

It searches, reads what comes back, and links the pages it used. OpenAI says that when ChatGPT search uses third-party search providers, it typically rewrites the question into one or more targeted queries for themOpenAI, and that its own crawler, OAI-SearchBot, is what surfaces websites in those answersOpenAI.

Which providers? OpenAI’s help pages have named Bing, as quoted by Semrush, and a Semrush test in August 2025 reported that Google results are used at least some of the timeSemrush. Treat both as reported, not as an official list. The practical reading is the same either way: be indexed in Bing and Google.

Two ways ChatGPT “knows” a business

What the model learned

Built into the model during training, from data that may include pages GPTBot crawled. It stops at a cut-off date, comes without links, and nothing you do today changes it on a schedule you control.

What search finds now

Fetched when someone asks, with links to the pages it read. This is where a small business can show up, and the part this guide is about: crawlable, indexed, clearly written pages, and other sites that mention you.

Can I pay for it?

No. OpenAI says ads in ChatGPT are labelled, shown beneath answers and do not influence the answer itself, as reported by NBC News in January 2026.

Do I have to submit my site?

No. OpenAI's publisher FAQ says any public website can appear in ChatGPT search; the condition it names is not blocking OAI-SearchBot.

Can anyone promise it?

No. Which pages an answer cites depends on the question, the search results and the model. Anyone who guarantees a mention is guessing.

Sources: NBC News OpenAI FAQ

Different job, same word: putting a ChatGPT-style assistant on your own site is a separate project, covered in how to train a chatbot on your website. This page is about your site showing up inside ChatGPT.

02 / Readiness check

Is your website ready
for ChatGPT search?

Tick what is true of your site today. You get a score, the three fixes worth doing first, a robots.txt block for the crawler policy you pick, and a regex that separates AI visits in GA4. It starts with a typical small business site. Nothing you tick leaves your browser.

ChatGPT readiness check · weights in brackets
What is true of your website

Your score

23/100 Start with the basics

Fix these first

  1. 1 robots.txt lets OAI-SearchBot in Allow OAI-SearchBot in robots.txt. OpenAI says a site that blocks it is left out of ChatGPT's search answers, apart from bare navigational links.
  2. 2 Your main pages are indexed in Bing and Google Check that your main pages are indexed: search site:yourdomain.com in Bing and in Google, and submit your sitemap in Bing Webmaster Tools and Google Search Console. ChatGPT search also asks outside search providers.
  3. 3 Other sites mention you: lists, reviews, press Get onto the lists, directories and review sites that come up when you search your own trade and town. Those are the pages an AI search reads when someone asks for a recommendation.

Your OpenAI crawler policy

Add to robots.txt 
User-agent: OAI-SearchBot
Allow: /

User-agent: GPTBot
Disallow: /

Your pages can appear and be linked in ChatGPT search, and OpenAI is asked not to use them for training. ChatGPT-User is left to your other rules; OpenAI says robots.txt may not apply to it, because a person asked for the page.

A crawler named in its own group ignores the rules under User-agent: *, so repeat any paths you also want it kept out of.

AI sources to count

GA4 regex for Source 
.*(chatgpt\.com|openai|perplexity|gemini\.google|copilot|claude\.ai).*

Use it as a “matches regex” condition on Source in a custom channel group, or on Session source in an exploration. The leading and trailing .* make it match anywhere in the source name. Steps are in chapter 07.

How the score works. Each item carries the weight in brackets: 3 for what can keep you out entirely, 2 for what an answer needs to say about you, 1 for housekeeping. The score is the share of the weight you have ticked, and the fixes are the heaviest unticked items, in list order. It is a checklist, not a prediction: a perfect score makes you eligible and quotable, not cited.

03 / OpenAI’s crawlers

Which OpenAI crawlers
should robots.txt allow?

Allow OAI-SearchBot if you want to appear in ChatGPT search. GPTBot is a separate choice about training, and blocking it does not block search. OpenAI controls the two with separate robots.txt settings and says a change takes about 24 hours to reach search resultsOpenAI.

OpenAI's crawlers and what each does, from OpenAI's crawler documentation, checked 29 October 2026
User agent What it does If you block it For this goal
OAI-SearchBot Surfaces websites in ChatGPT’s search features. Your pages are left out of ChatGPT search answers, though they can still appear as navigational links. Allow
GPTBot Crawls content that may be used to train OpenAI’s foundation models. Signals that your content should not be used for training. Search is unaffected. Your call
ChatGPT-User Visits a page when a person asks ChatGPT, or a custom GPT, about it. Not used for automatic crawling or to decide what appears in search. OpenAI says robots.txt rules may not apply, because a person started the visit. Leave it be
OAI-AdsBot Checks landing pages submitted as ChatGPT ads. Only visits pages submitted as ads. Only matters if you advertise in ChatGPT. Not needed

From OpenAI’s crawler overview, checked 29 October 2026OpenAI. OpenAI also publishes the IP ranges of each crawler there, in case your firewall or host blocks unknown bots.

Does my robots.txt block ChatGPT?

Open yourdomain.com/robots.txt, paste it here, and the tester reads it the way the robots standard describesRFC 9309:

  1. 01A crawler that has a group with its own name follows only that group, and ignores User-agent: *.
  2. 02Without its own group it follows the * group. With neither, it may crawl everything.
  3. 03The longest matching Allow or Disallow wins; on a tie, Allow wins. * matches anything and a final $ means “ends here”.

The example is a common WordPress file with GPTBot blocked. Notice that OAI-SearchBot is still allowed: it has no group of its own, so it follows *. Firewalls and “block AI bots” switches at your host can still turn a crawler away; this only reads the file.

robots.txt tester Runs in your browser
  • OAI-SearchBotChatGPT search

    Follows the User-agent: * group, and no rule there matches this path.

    Allowed
  • GPTBotTraining

    Follows its own group; the longest matching rule is Disallow: /.

    Blocked
  • ChatGPT-UserFetches a person asked for

    Follows the User-agent: * group, and no rule there matches this path.

    Allowed

ChatGPT-User is shown for completeness: OpenAI says robots.txt may not apply to it. Nothing you paste leaves this page.

04 / Answer-ready pages

How to make your pages
easy for ChatGPT to quote.

Write down, in plain text, the facts someone would ask about. An answer can only repeat what a page says. If your price lives in a PDF, your area on a map image and your services in a slideshow, there is nothing to quote.

  1. 01

    Who you are, what you do

    An about or home page whose first paragraph names the business, the service and the customer. "Gas-safe boiler repair and servicing for homes in Leeds" beats "Passionate about heating since 2009".

  2. 02

    Prices, or honest ranges

    Real numbers, "from" prices, or the range and what moves it. People ask what things cost, and a page without a figure can't be the one quoted.

  3. 03

    Where you work

    Your town, the areas you cover and whether you travel or ship, written as text.

  4. 04

    How to reach you

    A contact page with phone, email and hours in text. An answer that cites you should be able to send people somewhere useful.

  5. 05

    An FAQ from real questions

    The questions customers actually send you, answered in two or three sentences that make sense on their own, out of context.

  6. 06

    The same details everywhere

    Name, address and phone identical on your site, Google Business Profile, Bing Places and directories. When sources disagree, none of them is a safe thing to repeat.

None of this is special to ChatGPT. Google says the same about its own AI features: there is no special markup or file to add, and pages must be crawlable and indexed to be usedGoogle. A page written for a person who wants a quick, specific answer is the page an AI search can use.

05 / Mentions

How to get your website
recommended by ChatGPT.

Be on the pages a search returns for the question, not only on your own. “Who is the best X in Y?” is answered from search results, and for questions like that the results are usually lists, directories, review sites and local press. If you are on those pages, the answer has more than your own word to go on.

Step 01

Search your own question

Type the question a customer would ask into Bing and Google, the way they would say it, with your town. Note every list, directory and review site on the first page.

Step 02

Get onto those pages

Claim the directory listings, ask the list writers to consider you, and ask happy customers to review you where those sites collect reviews. Never pay for fake reviews or write your own.

Step 03

Keep the details matching

Every listing gets the same name, address, phone and a link to the page that best describes the service.

Step 04

Know which ones send people

Link each listing to your site with its own tagged link, so each one shows as its own source in your analytics.

For that last step, Builderify Short Links can add utm_source and the other UTM tags to the destination of each link, so a directory listing or a review profile carries its own name into your reports. Reviews are their own project: see how to get more Google reviews.

06 / llms.txt

Does llms.txt help you
get into ChatGPT?

Nobody has shown that it does. llms.txt is a proposal from Jeremy Howard, published in September 2024: a Markdown file at /llms.txt with a short summary of a site and links to its key pages, written for language models to readllmstxt.org.

OpenAI’s crawler documentation does not mention it, and we found no statement from OpenAI that ChatGPT search reads it. Google says its Search, AI features included, doesn’t use such files, and that keeping one neither helps nor harmsGoogle. So do the work in chapters 03 to 05 first.

If you add one anyway: the shape
# Northside Heating

> Boiler repair, servicing and installation for homes
> in Leeds and Wakefield. Repairs from £85.

## Services
- [Boiler repair](https://example.co.uk/boiler-repair): weekend call-outs
- [Prices](https://example.co.uk/prices)

## Contact
- [Contact and areas covered](https://example.co.uk/contact)

Cheap to add, and harmless. This site publishes one at /llms.txt, generated from the same page list as its sitemap; we keep it as a convenience, not because anyone has said it earns a citation.

07 / Track the visits

How to track traffic
from ChatGPT.

Look for the source chatgpt.com. OpenAI’s publisher FAQ says ChatGPT automatically adds utm_source=chatgpt.com to the links it shows, so ChatGPT referral traffic can be told apart in analyticsOpenAI. For where this fits among all your sources, see where website traffic comes from.

ChatGPT

Tags its links with utm_source=chatgpt.com, per OpenAI. A visit without the tag falls back to the referrer, or to direct.

Perplexity, Gemini, Copilot, Claude

We found no documented tag like ChatGPT's from any of them. Their visits arrive with the assistant's address as the referrer, when the browser or app passes one, or with nothing at all.

Nothing at all

Desktop and phone apps often send no referrer. Those visits land in direct traffic, and no tool can label them after the fact.

The tag can go missing. Around 8 to 15 October 2026, after the GPT-6 release, ChatGPT links lost the tag and the referrer, and those visits showed up as direct; Adobe’s Cedric Huesler reported on 15 October that it was a bug and the tags were back, as covered by Search Engine RoundtableSER. If your ChatGPT line dipped that week, annotate it rather than act on it. What direct traffic really is explains why untagged visits pile up there.

In GA4: an “AI assistants” channel

Google’s own help page uses AI assistants as its example of a custom channel groupGoogle. You need the Editor role or above.

  1. 1

    Go to Admin, then under Data display open Channel groups.

  2. 2

    Click Create new channel group. It starts as a copy of the default group, so every other channel keeps working.

  3. 3

    Click Add new channel, name it “AI assistants”, then Add condition group: Source, matches regex, and paste the regex from the readiness check. Save channel.

  4. 4

    Click Reorder and drag AI assistants above Referral, because GA4 gives each visit to the first channel it matches. Apply, then Save group.

Google says custom channel groups can be applied to reports retroactively, and a standard property can have two. For a one-off look instead, open Explore, add the Session source dimension and filter it with the same regex.

In Builderify: User Journey’s sources

User Journey names each visit’s source from its UTM tags first, before it looks at the referrer. So:

  • chatgpt.com A ChatGPT visit that carries utm_source=chatgpt.com is listed under the source chatgpt.com, exactly as the tag spells it.
  • your own tags A link you tag yourself, such as the one on a directory listing or review profile, shows under the utm_source you gave it.

Crawlers aren’t recorded as visitors, so what you see are people who clicked. Open a source to see its landing pages and the goals its visitors reached; to make those goals count enquiries, track your form submissions. Whether AI visits deserve a line in your weekly numbers is a question for the website metrics worth tracking.

Builderify can’t get you recommended by ChatGPT. Nothing can promise that. It can show you who arrives from it and what they do next. A tracked visitor journey costs 0.1 credits, every account gets 250 free credits a month, and one-time packs start at $9 with no monthly fee. See pricing.

Start free
Questions

ChatGPT and your website, before you ask.

No. OpenAI says ads in ChatGPT are labelled, shown beneath answers and do not influence the answer itself. What you can do is let OpenAI's search crawler, OAI-SearchBot, read your site, write pages that state what you do, where and for how much, and be listed on the directories and review sites a search turns up.

No. GPTBot is OpenAI's training crawler; ChatGPT search relies on OAI-SearchBot, which robots.txt controls separately. You can block GPTBot and allow OAI-SearchBot, and OpenAI says a robots.txt change takes about 24 hours to reach search.

OpenAI says ChatGPT adds utm_source=chatgpt.com to the links it shows, so look for the source chatgpt.com. In GA4, a custom channel group with an AI assistants channel, matched by a regex on Source and placed above Referral, collects ChatGPT, Perplexity, Gemini, Copilot and Claude visits that carry a source. Builderify User Journey reads UTM tags first, so those visits show under chatgpt.com.

Nobody has shown that you do. llms.txt is a 2024 proposal for a Markdown summary of a site; OpenAI's crawler documentation does not mention it, and Google says its Search does not use it. It is harmless to add, but crawler access, indexed pages and clear content come first.

No, and nobody can promise that. Builderify can show who arrives through ChatGPT's tagged links and what they do next: User Journey lists chatgpt.com as a source, with its landing pages and goals. A tracked journey costs 0.1 credits, and every account gets 250 free credits a month.

Sources

OpenAI’s crawler page, Google’s pages, llmstxt.org, Semrush and Search Engine Roundtable were read on 29 October 2026. OpenAI’s help-centre articles could not be loaded directly that day; their wording is as indexed by search engines on that date. How User Journey names a source comes from Builderify’s own code; prices come from the rate card on this site. The readiness score is our own weighting, not a measure of anything OpenAI uses.

← All guides

Stop losing the visitors
you already paid for.

More leads from the traffic you already have, and follow-up that sends itself. Free to start. No card, no monthly fee.

Start free See what it does