How to Get Your Website Cited by ChatGPT (2026)

To get your website cited by ChatGPT, it first has to be reachable and readable. Allow OpenAI's search crawler (OAI-SearchBot) in robots.txt and in your CDN or firewall, get your pages indexed in the search engines ChatGPT draws on, and serve the main text as plain HTML. Then publish pages that answer specific questions clearly. Nobody can guarantee a citation, but you can remove everything that rules you out.

ChatGPT only links to sources when it searches the web. Answers from its trained knowledge often name brands without linking to anyone. This guide covers both cases, sticks to what OpenAI actually documents, and says plainly where documentation stops.

Key takeaways

• Search citations depend on OAI-SearchBot. OpenAI says sites that opt out of it won't be shown in ChatGPT search answers. GPTBot (training) is a separate choice.
• Blocks often sit outside robots.txt. CDN bot rules, firewalls and security plugins can stop OpenAI's crawlers even when robots.txt allows them.
• Being indexed in search engines helps, but rankings don't map neatly to citations. Studies disagree a lot on how closely ChatGPT's citations match Bing or Google results.
• Text that only appears after JavaScript runs is a risk. One crawler study found OpenAI's crawlers don't execute JavaScript.
• Measure it. Ask the questions your customers ask, check referrals from chatgpt.com, and look for OAI-SearchBot in your server logs.

How does ChatGPT choose which websites to cite?

ChatGPT cites websites when it runs a web search for a prompt. It pulls pages, writes an answer and links the sources it used. When it answers from what it learned in training, there's usually no link. OpenAI doesn't publish how it ranks or picks sources, so any "ChatGPT ranking factor" list is guesswork.

What we do know comes from OpenAI's crawler documentation and from outside studies. The OpenAI crawler overview says OAI-SearchBot is "used to surface websites in search results in ChatGPT's search features." So the search path depends on crawling.

Outside studies try to reverse-engineer the rest, and their results don't agree:

• A Seer Interactive study (February 2025, about 100 questions) found that 87%+ of ChatGPT search citations matched Bing's top 20 organic results and 56% matched Google's. The authors call the results "directional" because the sample was small.
• An Ahrefs study (15,000 long-tail queries, July 2025) found only about 8% of ChatGPT's in-text citations ranked in Google's or Bing's top 10 for the original prompt. It suggests AI assistants break a prompt into several related searches ("query fan-out") instead of copying a single results page.

The practical lesson: being indexed and visible in search engines is a minimum requirement, but ranking for your main keyword doesn't decide whether you get cited. Pages that answer the narrower sub-questions behind a prompt get their own chances.

Can ChatGPT crawl your website?

ChatGPT can only cite pages its search crawler is allowed to fetch. Check that robots.txt doesn't block OAI-SearchBot, and that your CDN, firewall or hosting doesn't block OpenAI's published IP ranges. This is the most common fixable reason a site never shows up, and it often appears after a security change nobody linked to AI search.

OpenAI runs three separate user agents, and each one is controlled on its own:

Source: OpenAI crawler overview, as of November 2026.

A robots.txt that allows search citations but opts out of training looks like this:

User-agent: OAI-SearchBot
Allow: /

User-agent: GPTBot
Disallow: /

OpenAI notes it can take about 24 hours after a robots.txt change before its systems adjust. Our robots.txt guide covers the syntax and the most common mistakes, such as a broad User-agent: * disallow left over from a staging site.

Next, check everything that sits in front of your server. OpenAI publishes the IP ranges for OAI-SearchBot at openai.com/searchbot.json, so you can allowlist them or check your logs against them. If you use Cloudflare, its AI crawler and agent access policies let you allow or block Search, Agent and Training traffic separately. Make sure the "Search" setting matches what you actually want. Security plugins and hosting-level bot protection can return 403 errors to crawlers without anyone noticing.

Does your site need to rank in Bing to get cited by ChatGPT?

You don't need a top Bing ranking, but being indexed in Bing and Google is a sensible baseline. One study found ChatGPT's citations closely matched Bing's results, while a larger one found little overlap with either engine's top 10. Neither study is official, and OpenAI doesn't document a dependence on any one search engine.

Being indexed costs little, so cover it anyway:

1. Verify your site in Bing Webmaster Tools and submit your XML sitemap there, as you would in Google Search Console.
2. Use IndexNow if your CMS or CDN supports it. The IndexNow FAQ says it lets you notify Bing and other participating engines when you add, update or delete URLs, and also that "submitting a URL does not guarantee immediate indexing."
3. Fix the usual indexing problems first. Noindex tags, wrong canonicals, thin pages and orphan pages keep you out of every index, AI or not.

Because of query fan-out, think beyond your main keyword. A prompt like "best accounting software for a three-person agency" may trigger searches for pricing, integrations and comparisons. A clear pricing page, a comparison page and a well-organized integrations page each give ChatGPT something to cite.

Can ChatGPT read the content on your pages?

ChatGPT can only quote text its crawler receives. If your main content, prices or product details appear only after JavaScript runs, it may never see them. OpenAI doesn't document whether OAI-SearchBot renders JavaScript, but one independent crawler study found it does not.

Vercel's analysis of AI crawler traffic (December 2024) found that "none of the major AI crawlers currently render JavaScript." OpenAI's crawlers downloaded JavaScript files but didn't execute them. Vercel recommends server-rendering any important content.

Here's a quick test. Open a key page, view the page source (not the inspector, which shows the page after JavaScript has run) and search for a sentence from your main content. If it isn't in the raw HTML, a crawler that doesn't render JavaScript gets an empty shell. This often happens with:

• single-page apps that render everything in the browser
• tabs and accordions whose contents load on click
• prices, stock levels and reviews pulled in by scripts
• "load more" lists where later items never reach the HTML

The fixes are server-side rendering, static generation or pre-rendering for important templates. Our guide to JavaScript rendering problems explains how to test and choose between them.

What kind of content does ChatGPT cite?

OpenAI hasn't published content guidelines for ChatGPT citations. What works in practice is the same thing that makes any page easy to quote: a direct answer near the top, specific facts, clear headings, and a page focused on one question instead of ten.

Since selection criteria aren't documented, treat these as good editorial habits, not guaranteed ranking factors:

• Lead with the answer. Answer the question in the first 40–60 words under a heading phrased as that question, then explain.
• Write passages that make sense on their own. An AI system may quote one paragraph without the rest of the page. "Our plan includes this" makes no sense out of context. "The Basic plan includes three users" does.
• Be concrete. Name products, places, specs, limits and conditions. Vague marketing copy gives nothing to quote.
• Cover the sub-questions. Comparison, pricing, "how does it work" and "is it right for X" pages match the narrower searches that fan-out produces.
• Keep facts current and consistent. If your pricing page, an old blog post and a directory listing disagree, any system pulling from them can repeat the wrong one.

Our guide to answer engine optimization shows how to structure answer-first pages in detail.

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

Without a web search, ChatGPT answers from what it learned in training, so mentions depend on how often and how consistently your business is described across the web. You can't submit your site for this. You can make sure accurate, consistent information about your business exists in many places.

Practical steps, none of which OpenAI documents as a citation signal:

• Decide on GPTBot deliberately. Blocking it signals that your content shouldn't be used for training. That's a legitimate choice, but it's a trade-off if you want models to know about you. It doesn't affect search citations either way.
• Keep your entity data clean. Use Organization structured data with sameAs links to your official profiles, and a correct Wikidata entry if your business qualifies for one. They give machines one consistent description of who you are.
• Earn mentions on independent sites. Reviews, industry directories, press coverage and comparison articles describe your business in words you don't control. The Seer study noted that citations from affiliates, news sites and aggregators were common for "best/top" questions.
• Make your own "about" facts easy to find. Founding year, location, what you sell, who it's for, in plain text on a crawlable page.

How do you check whether ChatGPT cites your website?

Use three signals: ask ChatGPT the questions your customers ask and note which sites it cites, watch your analytics for referrals from chatgpt.com, and check your server logs for OAI-SearchBot visits. Together they show whether you're crawled, cited and getting clicks.

1. Prompt testing. Write 20–30 real customer questions, such as "who installs heat pumps in Leeds" or "best invoicing tool for freelancers in Spain". Run them with search turned on, record the cited domains and repeat monthly. Answers vary between runs, so look at trends, not single answers.
2. Referral traffic. In your analytics, filter sessions where the source is chatgpt.com, and check which landing pages they reach.
3. Server logs. Count OAI-SearchBot requests and their status codes, and check the IPs against OpenAI's published ranges. A crawler that keeps getting 403 or 5xx errors explains a lot. See our guide to server log analysis.

To check the technical side of your own site, a free SEORecheck audit looks at AI crawler access in robots.txt (search, user and training bots), snippet controls, text visible only after JavaScript, Organization/sameAs markup and your Wikidata entry, and how your site answers the questions people ask AI assistants about your business. You can see how these findings are reported in the sample report.

For a broader set of AI and classic search metrics, see SEO vs GEO performance measurement.

What doesn't help you get cited by ChatGPT?

Paid "guaranteed ChatGPT placement" services, keyword-stuffed pages and markup added only for AI don't have documented support. OpenAI doesn't offer a submission form or a paid way into organic citations. Spend your time on access, indexing, readable HTML and clear answers.

ChatGPT citations reward the same groundwork as classic search: a site crawlers can reach, pages that are indexed, content in the HTML, and answers specific enough to quote. Get those right, then track the results over months, not days.