What does ChatGPT actually fetch from my site?

ChatGPT's crawlers behave like a fast, headless reader: they fetch the raw HTML, follow same-origin links shallowly, and read robots.txt, sitemap.xml, llms.txt, and ai.txt if present. They do not execute most JavaScript. If your content only renders client-side, the bot sees an empty shell.

How do I verify ChatGPT is citing my page?

Filter your analytics for referrer hosts chatgpt.com and chat.openai.com, and add a segment for the user-agent string ChatGPT-User. Cross-check by typing the exact query into ChatGPT with browsing enabled and looking at the citation list. Track week over week in a citation log.

Why is my page not showing up in ChatGPT?

The four most common reasons are: robots.txt disallows OAI-SearchBot or GPTBot; your content is rendered by client-side JavaScript and the bot sees an empty page; your first paragraph does not directly answer the query; or your page has no canonical URL, no last-modified date, and no quotable passage the model can attribute.

Writing

How to show up in ChatGPT results (2026 guide)

Q: Which crawlers does ChatGPT use?

ChatGPT uses three OpenAI bots. OAI-SearchBot indexes pages for the in-product search citation feature. ChatGPT-User fetches a page on demand when a user clicks a citation or when ChatGPT browses live. GPTBot crawls pages that may be used to improve future models. Allow OAI-SearchBot and ChatGPT-User in robots.txt at minimum; allow GPTBot if you want training inclusion.

Last updated: April 18, 2026

Short answer

ChatGPT cites pages that are (1) crawlable by OAI-SearchBot and GPTBot, (2) answer the user's exact question in the first paragraph, and (3) expose a clean, citation-ready surface — a canonical URL, a last-modified date, and ideally a hand-curated llms.txt or a markdown endpoint. Allow the bots, lead with the answer, and give the model a stable thing to attribute.

Which crawlers does ChatGPT use?

OpenAI publishes three crawlers. Each has a different job, and many sites accidentally block the wrong one.

OAI-SearchBot — indexes pages for citation inside ChatGPT's search/browse feature. This is the one you most need to allow.
ChatGPT-User — fetches a single page on demand when a ChatGPT user clicks a citation or asks the model to browse live. Treat this like a real visitor.
GPTBot — crawls pages that may be used to improve future OpenAI models. Allowing this affects training, not retrieval.

If you want to be cited but opt out of training, allow OAI-SearchBot and ChatGPT-User, and disallow GPTBot. If you want both, allow all three. Disallowing all three means ChatGPT will paraphrase your content from cached sources without attributing it to you.

What does ChatGPT actually fetch?

The crawlers behave like a fast, headless reader. They fetch raw HTML, read robots.txt and sitemap.xml, and pick up llms.txt and ai.txt if present. They do not execute most JavaScript. If your homepage relies on client-side rendering to display its main content, the bot sees an empty shell — and the model has nothing to cite.

The cheapest test:

curl -A "OAI-SearchBot/1.0" -L https://your-site.com/ | wc -c

If that number is a few hundred bytes, ChatGPT effectively cannot read your homepage. You either need server-side rendering, prerendering for known bot UAs, or a parallel plain-text surface like /shell.md.

The 4 things to ship this week

Allow the bots in robots.txt. Add an explicit User-agent block for OAI-SearchBot, ChatGPT-User, and (optionally) GPTBot. Don't rely on User-agent: * — many CDNs and security plugins quietly block AI bots that don't have a named allow rule.
Lead with the answer. Every page that targets a question-shaped query should have a one-paragraph direct answer in the first 200 words. That paragraph is what the model lifts. The rest of the page is supporting evidence.
Make the page easy to attribute. Set <link rel="canonical">, return a real Last-Modified header, and add JSON-LD Article + FAQPage blocks. Models prefer to cite pages where the URL, title, and freshness are unambiguous.
Publish an llms.txt. A short, hand-curated map of your most citeable URLs (see our llms.txt guide). It is not a magic bullet, but it is a strong hint to the retriever about which of your pages are canonical sources.

How do I verify ChatGPT is citing me?

There is no Search Console for ChatGPT yet, so you have to instrument it yourself. Three signals to watch:

Referrer hosts — segment your analytics by referrer containing chatgpt.com or chat.openai.com. These show up when a user clicks a citation in the ChatGPT UI.
User-agent log filter — in your access logs, search for ChatGPT-User, OAI-SearchBot, and GPTBot. Rising on-demand ChatGPT-User hits to a specific URL is a strong signal that URL is being cited.
Manual check — type the exact target query into ChatGPT with browsing enabled and read the citation list. Do this weekly for your top 10 queries and log it (we keep /reports for this).

Common reasons you're invisible

JavaScript-only rendering. The bot reads HTML, not your React tree. Server-render the answer paragraph at minimum.
Cloudflare / WAF blocking. Many security defaults block unknown bot UAs. Check your firewall logs for 403s on OAI-SearchBot and add an explicit allow.
The answer is buried. If the model has to read 3,000 words to find the answer, it will lift from a competitor whose answer is in the first paragraph.
No stable URL. Pages that change slug frequently or live behind tracking-parameter-only URLs are hard to attribute and deprioritized.
You're invisible to Bing too. ChatGPT's search backend historically leans on Bing's index. If Bing can't find you, ChatGPT often can't either. Submit a sitemap to Bing Webmaster Tools.

Next steps

The fastest way to know where you stand is to run our free audit — it fetches your homepage as each major AI crawler and tells you which ones see real content vs. an empty shell. Then read the companion posts:

Audit my site

Sources

Cite this page

title: How to show up in ChatGPT results (2026 guide)
url: https://pseudobash.com/blog/show-up-in-chatgpt-results
last-modified: 2026-04-18
snippet: ChatGPT cites pages that are crawlable by OAI-SearchBot and GPTBot, answer the question in the first paragraph, and expose a stable citation surface (canonical URL, Last-Modified, llms.txt).
X-Pseudobash-Cite: title="How to show up in ChatGPT results (2026 guide)"; url="https://pseudobash.com/blog/show-up-in-chatgpt-results"; last-modified="2026-04-18"