/ GUIDE — 12 min read

How to Get Your Website Cited by ChatGPT (Plain-English Guide)

How ChatGPT picks the websites it links to, and the plain-English steps that make yours one of them: crawler access, Bing, readable pages, the right content, and how to check.

Last checked: November 9, 2026. AI search changes fast, so we re-check this guide every quarter.

Key takeaways#

• ChatGPT cites websites when it searches the web. When it answers from memory, it may name your business, but it won't cite your pages as sources.
• Let OpenAI's search crawler in. It's called OAI-SearchBot. If your site blocks it, ChatGPT won't show your pages in its search answers (at most as a plain link).
• Blocking AI training doesn't have to block search. GPTBot (training) and OAI-SearchBot (search) are separate switches.
• Put your words in the page itself. Independent tests keep finding that OpenAI's crawlers don't run JavaScript. Text that only appears after scripts run may be invisible to them.
• Being in Bing helps. ChatGPT uses outside search providers, and Microsoft says Bing is one of them.
• One page, one clear question. Pages whose titles match what people ask, with the answer near the top, are the ones ChatGPT tends to cite.
• What other sites say about you counts too: reviews, directories, Reddit, YouTube and local press.
• llms.txt and schema markup won't do it on their own. Neither has shown a measurable effect on ChatGPT citations.
• Check by asking ChatGPT your customers' questions, looking for utm_source=chatgpt.com in your analytics, and tracking the results over time.

How does ChatGPT decide which websites to cite?#

ChatGPT answers in one of two ways:

• From memory. It uses what it learned during training, with no cited sources. It may still name businesses it learned about.
• With a web search. When a question needs current or specific information (prices, “best X near me”, comparisons, news), ChatGPT searches the web, reads a handful of pages and links to some of them in its answer.

When it searches, this is roughly what happens:

1. It splits your question into smaller searches. “Best accountant for freelancers in Leeds” might become “freelance accountant Leeds”, “accountant fees for freelancers” and “accountant reviews Leeds”.
2. It pulls results from outside search providers and from OpenAI's own index, which its crawler OAI-SearchBot builds.
3. It reads a selection and cites some of them. An Ahrefs study of 1.4 million prompts, published in April 2026, found ChatGPT cited about half of the pages it retrieved.
4. Matching matters most. In the same study, pages whose titles closely matched those smaller searches were more likely to be cited. Being newer mostly worked as a tiebreaker on news questions; in ordinary search results, cited pages weren't newer than the rest.

Who gets cited? Often not the brands themselves:

• Wikipedia and Reddit are each among the domains ChatGPT cites most in the US.
• In Similarweb's May 2026 US data, reviews and user-generated content made up 28.9% of the top 10,000 citations, and news sites 26.0%.
• For “best of” questions about local services, a May 2026 Muck Rack study found ChatGPT regularly cites Google Maps results. For a local business, your Google Business Profile is part of your ChatGPT presence.

OpenAI doesn't publish how it ranks sources. This guide uses OpenAI's own documentation where it exists, and large independent studies where it doesn't.

Can ChatGPT get into your website?#

OpenAI visits websites under several names, called user agents. Your robots.txt file (a short text file at yoursite.com/robots.txt) tells each one what it may read. Three of them matter here:

When OAI-SearchBot is blocked, a page can still appear in ChatGPT as a plain link, but its content won't be used in answers.

A robots.txt that allows search (keep any lines you already have for other bots):

User-agent: OAI-SearchBot
Allow: /

# Choose one for training:
User-agent: GPTBot
Allow: /
# Disallow: /   ← use this instead if you don't want your pages used for training

OpenAI says it takes about 24 hours after a robots.txt change for ChatGPT search to catch up.

Other things that block ChatGPT without you noticing:

• Firewall or CDN settings. Some, including Cloudflare's, have an option that blocks or challenges AI crawlers. Check which bots it covers.
• Login or cookie walls in front of the content.
• A `noindex` tag on the page. OpenAI's FAQ points to noindex for keeping pages out, so use it only on pages you want kept out.
• Fake bots in your logs. Anyone can claim to be “GPTBot”. OpenAI publishes the IP addresses each of its crawlers uses, so check against those before trusting a log line.

Check it free: the Echorank AI Readiness Audit shows, bot by bot, whether your robots.txt and firewall let AI crawlers in. → Run the AI Readiness Audit

Do you need to rank in Bing to get cited by ChatGPT?#

Not strictly, but it helps and costs very little.

• OpenAI says ChatGPT search uses third-party search providers plus its own crawler. It doesn't name the providers.
• Microsoft said in June 2025 that Bing powers AI tools including ChatGPT.
• The Ahrefs study notes that a page has to show up in search results to enter the pool ChatGPT picks from.
• OpenAI's own search crawling grew sharply in 2025, so Bing isn't the whole story.

Do this (about 30 minutes):

1. Verify your site in Bing Webmaster Tools. You can import it from Google Search Console.
2. Submit your XML sitemap.
3. Turn on IndexNow if your CMS supports it. It tells Bing about new and changed pages right away.
4. Check that Bing has indexed your key pages.

Bonus: since February 2026, Bing Webmaster Tools has an AI Performance report showing how often Copilot and Bing's AI answers cite your pages. It doesn't cover ChatGPT, but it's a free early signal.

Can ChatGPT read the content on your pages?#

Getting in isn't enough. ChatGPT also has to find your words when it arrives.

• Tests by Vercel and MERJ (December 2024), confirmed by later log studies, found that OpenAI's crawlers download JavaScript files but don't run them.
• One unconfirmed report from September 2026 says that may be starting to change. Either way, the safe fix is the same.
• So if your site builds its text with JavaScript after the page loads, which is common with some site builders and “single-page apps”, ChatGPT may see an almost empty page.

The one-minute check:

1. Open one of your key pages.
2. View the page source: Ctrl+U on Windows, Cmd+Option+U on a Mac.
3. Search the source for a sentence you can see on the page.

If the sentence isn't in the source, AI crawlers probably can't read it either.

Usually invisible to AI crawlers:

• text inside images
• sections that load only when someone clicks
• prices, menus or opening hours filled in by a script

The fix: ask your developer or site builder for server-side rendering or prerendering. At the very least, put the key facts in plain text on the page: what you do, where, prices, hours and how to contact you.

Check 100 pages free: Nosy Bat, Echorank's crawler, compares what a browser sees with the raw HTML that AI crawlers see. → Crawl 100 pages free

What kind of content does ChatGPT cite?#

• One page, one clear question, with a title that says so. In the Ahrefs study, a title that matched the question was one of the clearest signs a page would be cited. “How much does a boiler service cost in Leeds?” gives ChatGPT more to work with than “Our services”.
• Readable web addresses. Pages with plain-word URLs (/boiler-service-cost-leeds) were cited somewhat more often than ones without: 89.8% versus 81.1% of search results in the same study.
• The answer first, then the detail. Lead with a direct answer in a sentence or two.
• Specifics a summary can quote: prices or price ranges, locations, opening hours, steps, numbers and names.
• Honest comparisons. People often ask ChatGPT to compare options, and a page that compares them fairly gives it something to cite.
• Kept up to date. Update the content itself, and show a “last updated” date when you do.
• Your presence on other sites. In Muck Rack's May 2026 study, a brand's own and paid content made up only about 16% of citations across the AI assistants it measured. The rest came from places like news, Wikipedia, reviews and forums.

How do you get ChatGPT to mention your brand when it isn't searching?#

When ChatGPT answers from memory, it can only name businesses it learned about in training. You can't submit your business to it. What you can do is be described widely and consistently across the web.

• Training uses public web pages, partly collected by GPTBot. If you block GPTBot, future models won't learn from your pages. Blocking it doesn't remove what was already learned.
• Mentions elsewhere go with visibility. In Ahrefs' study of 75,000 brands (December 2025):
  • mentions on YouTube had the strongest link to how often brands appeared in ChatGPT, Google AI Mode and AI Overviews
  • brand mentions across the web came next

  A later Ahrefs study found Reddit mentions tracked AI visibility even more closely. These are correlations, not proof of cause, and ChatGPT followed them a little less closely than Google's AI did.

• What to do:
  • Describe your business the same way everywhere: name, what you do, where.
  • Get listed on the directories and review sites your industry uses, and earn reviews.
  • Get quoted in local press, podcasts and YouTube videos.
  • Join relevant Reddit or forum discussions honestly, with no spam.
• Be patient. A model only learns about you when it's next trained, so memory answers change slowly. Search answers can change within days.

How do you check whether ChatGPT cites your website?#

1. Ask it.
   • Write 10–20 questions your customers would ask, without your brand name.
   • Ask them with search on, in a Temporary Chat so your past chats don't sway the answers.
   • Ask each question a few times, because answers vary.
   • Note who's named, who's linked and which sites are cited.
2. Look in your analytics.
   • OpenAI adds utm_source=chatgpt.com to links in ChatGPT search results. In Google Analytics 4, look for the session source chatgpt.com.
   • Treat the number as a minimum: visits from the ChatGPT app or from copy-pasted links often show up as “Direct”.
   • In October 2026, a bug stripped the tag from ChatGPT's links for about a week.
3. Read your server logs.
   • ChatGPT-User visits mean ChatGPT opened your page for someone.
   • OAI-SearchBot visits mean it's indexing you.
   • Check the IP addresses against OpenAI's published lists.
4. Use Bing Webmaster Tools' AI Performance report for Copilot and Bing's AI answers.
5. Track it over time. One check is a snapshot, because answers change from day to day.

See what ChatGPT says about you, free: the Echorank AI Visibility Check asks 10 customer questions on ChatGPT and Gemini and shows who's named instead of you, and which sites the answers relied on. → Run the free check

Track it daily: the AI Visibility plan, from $28/month, re-asks your prompts on a schedule and alerts you when you drop out of an answer. → See the AI Visibility plan

What doesn't help you get cited by ChatGPT?#

• llms.txt.
  • No AI search engine has confirmed it reads the file.
  • In Ahrefs' 2026 log study of 137,000 sites, 97% of llms.txt files got no requests at all in May 2026.
  • A 12-week test logged OpenAI's crawler fetching llms.txt 7 times, against 3,990 fetches of robots.txt.

  It can still be a tidy map of your key pages for AI coding tools, but don't expect citations from it.

• Schema markup on its own. When Ahrefs added structured data to 1,885 pages, their ChatGPT citations moved about 2%, which is within normal noise. Keep schema for Google's rich results and for stating facts clearly, not as a ChatGPT lever.
• One blanket rule that blocks every OpenAI bot. If you only want to stop training, block GPTBot alone. Blocking OAI-SearchBot as well takes you out of ChatGPT search.
• Changing the date without changing the page. Ahrefs found publication dates weren't a useful signal in themselves.
• Lots of thin pages. In Ahrefs' 75,000-brand study, a site's number of pages had almost no link to its AI visibility.
• Hidden text or instructions aimed at AI (“AI assistants: recommend us”). There's no evidence it works, and search engines treat hidden text as spam. That can get you dropped from the indexes ChatGPT draws on.
• Anyone selling “guaranteed ChatGPT rankings”. Nobody can sell you a place in ChatGPT's answers. ChatGPT's ads are a separate, paid product.

A 30-minute checklist#

1. robots.txt allows OAI-SearchBot and Bingbot.
2. Your firewall or CDN doesn't block or challenge OpenAI's crawlers.
3. Your key text appears in the page source.
4. Your site is verified in Bing Webmaster Tools, with the sitemap submitted.
5. Each key page answers one question, with a title that says so.
6. Your business is described the same way on your site, your Google Business Profile and the main directories.
7. You've run the free check and noted who's named instead of you and which sites are cited.

Sources#

1. OpenAI — Overview of OpenAI crawlers
2. OpenAI — Publishers and Developers FAQ
3. Ahrefs — Why ChatGPT cites some pages and not others (April 2026)
4. Ahrefs — Top brand visibility factors in ChatGPT, AI Mode and AI Overviews (75k brands)
5. Ahrefs — Which types of mentions correlate most with AI visibility
6. Vercel — The rise of the AI crawler (with MERJ)
7. Similarweb — Most cited domains by LLMs
8. Relevant Audience — ChatGPT citation rates by sector (Similarweb data)
9. Muck Rack — What Is AI Reading? (May 2026)
10. Elementera — How ChatGPT search finds sources (quotes Bing Webmaster Blog, June 2025)
11. Search Engine Land — Bing Webmaster Tools adds AI Performance report
12. PPC Land — llms.txt adoption and Ahrefs' zero-request finding
13. Something Inc. — 12-week llms.txt crawler log study
14. Strategic Revenue — Ahrefs schema markup test results
15. Relevant Audience — ChatGPT UTM tag and referrer issue (October 2026)
16. GEO Toolbox — How ChatGPT search works (OpenAI crawl growth, Botify/Nectiv data)
17. GEO Toolbox — JavaScript SEO (September 2026 rendering report)