GPTBot vs OAI-SearchBot: which to block?
Short answer: GPTBot collects training data and OAI-SearchBot decides whether ChatGPT can cite you. Block GPTBot if you want out of training. Leave OAI-SearchBot alone unless you want out of ChatGPT’s answers too. Registry verified 19 September 2026.
What’s the difference between GPTBot and OAI-SearchBot?
OpenAI runs three tokens that people meet in a robots.txt, and each one does a different job. The third, ChatGPT-User, is not a crawler at all: it fetches one page when a person asks ChatGPT to look at it.
| GPTBot | OAI-SearchBot | ChatGPT-User | |
|---|---|---|---|
| Its job | Model training | Search & citations | Fetches when a user asks |
| Blocking it affects | OpenAI model training | ChatGPT search answers and citations | ChatGPT opening your page when someone asks it to (a request, not a control) |
| Blocking it leaves alone | ChatGPT search answers and citations | OpenAI model training | ChatGPT search results; OpenAI model training |
| Obeys robots.txt | Yes, per its operator's docs | Yes, per its operator's docs | May bypass, per its operator's docs |
| Verifiable by IP | Yes, ranges are published | Yes, ranges are published | Yes, ranges are published |
| Last verified | 2 September 2026 | 2 September 2026 | 2 September 2026 |
Does blocking GPTBot remove you from ChatGPT?
No. GPTBot gathers pages to train future OpenAI models. The answers ChatGPT gives when it searches the web come from a different crawler, so a GPTBot block keeps you out of training and leaves your citations exactly where they were. It is the one OpenAI rule you can write with no visibility cost.
What happens if you block OAI-SearchBot?
ChatGPT stops being able to surface and cite your pages when somebody asks a question you have answered. OpenAI notes that a blocked site can still turn up as a bare navigation link, which is a name with nothing behind it and not the same as being recommended.
This is the costly mistake, and it rarely looks like one. A “block AI scrapers” snippet lists every OpenAI token its author could find, it gets pasted in with good intentions, and the site leaves ChatGPT’s results with no error anywhere. See the OAI-SearchBot page for the details.
Where does ChatGPT-User fit?
It runs when a person pastes your URL into ChatGPT or a GPT reaches for your page. OpenAI says robots.txt may not apply to these user-initiated fetches, so a rule for ChatGPT-User records a preference rather than enforcing one. If a page must not be read by anyone holding the link, robots.txt is the wrong layer for it.
What should your robots.txt say?
For most sites that want out of training and into ChatGPT’s answers, this:
User-agent: GPTBot
Disallow: /
User-agent: OAI-SearchBot
Allow: /The explicit Allow is worth having even though allowing is the default: it keeps OAI-SearchBot in if somebody later adds a wildcard Disallow for another reason. If you run ChatGPT ads, leave OAI-AdsBot out of any block list too, or your own landing pages fail review.
Already have a file? Paste it into the checker to see which of the three it blocks. Starting fresh? The generator writes this for you.
Can a correct robots.txt still keep ChatGPT out?
Yes. A CDN or bot-management rule can turn OAI-SearchBot away before it ever reads the file, and nothing in robots.txt will show it. That is the most common reason a site with a perfect file is still missing, and Cloudflare is the usual one. The scan requests your home page as each crawler and shows where the answers differ.
We request your robots.txt and homepage. We read your sitemap only when robots.txt hides pages from a crawler and we need to measure how much is affected. No other page.
Crawlers in this guide
- GPTBot: Model training, run by OpenAI
- OAI-SearchBot: Search & citations, run by OpenAI
- ChatGPT-User: Fetches when a user asks, run by OpenAI
Where to go next
- ClaudeBot vs Claude-SearchBot: which to block?
Anthropic's crawlers split the same way OpenAI's do, and the legacy tokens in most files do nothing at all. - Why your site is not showing up in ChatGPT
The common failure modes, in order, with a check you can run for each one. - Is Cloudflare blocking AI crawlers?
A correct robots.txt can still leave a site inaccessible. Here is what to look for when Cloudflare is in the way.