How to block GPTBot
(and when you should not)
GPTBot is OpenAI's training crawler. Here are the lines to block it, what a block does and does not do, and why blocking it is not the same as blocking ChatGPT.
What GPTBot is
OpenAI sends several automated visitors to websites. GPTBot is the one that collects pages for training. OpenAI says GPTBot crawls content that may be used to train its generative AI foundation models. It also says that disallowing GPTBot tells OpenAI a site's content should not be used for that training.
GPTBot is one of three OpenAI names you will see in robots.txt files.
- GPTBot: training. Collects content that may be used to train OpenAI's models.
- OAI-SearchBot: search. Finds pages to show and link to in ChatGPT's search features.
- ChatGPT-User: user fetches. Visits a page when a person using ChatGPT asks for something that needs it.
Each one is controlled separately. Blocking one does not block the others. For a side-by-side view, see OAI-SearchBot vs GPTBot vs ChatGPT-User.
The robots.txt lines
robots.txt is a plain text file at the root of your site. It must open at yourbusiness.com/robots.txt. If you do not have one, create a text file called robots.txt and upload it there.
Block GPTBot from your whole site
User-agent: GPTBot
Disallow: /That is all it takes. These two lines ask GPTBot not to read any page on the site.
Block training, but stay open to ChatGPT search
User-agent: GPTBot
Disallow: /
User-agent: OAI-SearchBot
Allow: /
User-agent: *
Allow: /The OAI-SearchBot group is not strictly needed if nothing else blocks it. Naming it makes your intent clear to anyone who reads the file later, including you.
Block GPTBot from part of your site
User-agent: GPTBot
Disallow: /blog/
Disallow: /resources/This keeps GPTBot out of those two folders and lets it read everything else.
Mistakes that undo the rule
- Forgetting your other rules. A crawler follows only the one group that best matches its name. Once you add a GPTBot group, GPTBot ignores your
User-agent: *group. Copy any rules you still want it to follow into its own group. - A misspelt name. A typo matches nothing, so the rule does nothing.
- The wrong host. Each host needs its own file. shop.yourbusiness.com does not read the file on yourbusiness.com.
- Not checking the live file. After any change, open yourbusiness.com/robots.txt in a browser and read what is actually there.
What blocking GPTBot does
- It states your choice on training. It asks OpenAI not to use pages GPTBot crawls for training its models.
- It separates training from search. OpenAI says that if a site allows both GPTBot and OAI-SearchBot, it may use the results of one crawl for both. A GPTBot block is how you split the two.
- It is public. Anyone can read your robots.txt, so your position is on record.
What blocking GPTBot does not do
- It does not take you out of ChatGPT search. That is OAI-SearchBot's job, and it has its own rule.
- It does not stop ChatGPT-User. OpenAI says ChatGPT-User acts when a user asks, and that robots.txt rules may not apply to it.
- It does not reach back in time. robots.txt tells crawlers what to fetch from now on. It is not a request to delete anything collected before.
- It does not block other companies. ClaudeBot, CCBot, Meta-ExternalAgent and others each need their own lines. CCBot builds a public archive of the web that others use, including to train AI models. See the AI crawler list.
- It is not a lock. robots.txt is a request. Well-behaved crawlers follow it, but it does not physically stop anyone. OpenAI publishes the IP addresses its crawlers use, so your host or developer can check that a visitor calling itself GPTBot really is.
When you should not block GPTBot
There is no right answer for every business. Blocking GPTBot is a reasonable choice if you do not want your writing, photos or product descriptions used to train models. But stop and think if one of these applies.
- You really meant to block ChatGPT search. GPTBot is the wrong name for that. And if your goal is to be found, blocking OAI-SearchBot or ChatGPT-User would work against you.
- You are pasting a list from somewhere else. Many "block AI" lists include search crawlers and user fetchers too. Check each name before you copy it. The AI crawler directory has a page for each one.
- Something already blocks it. Some hosts and CDNs block AI crawlers before robots.txt is even read. Find out where any existing block comes from before you add another. See Cloudflare's AI bot blocking and where blocks hide.
- You have not weighed what your pages are for. A publisher whose writing is the product may see training very differently from a shop whose pages mainly describe what it sells. Neither view is wrong. Decide on purpose.
Check what your site says now
Our free AI crawler checker reads your robots.txt and shows, name by name, whether GPTBot and 13 other AI crawlers are allowed. It also asks your homepage for a page as GPTBot, to see if your server turns it away. It checks the rules your site publishes. It does not test what OpenAI actually does.
For GPTBot's own page, with the lines to allow or block it, see What is GPTBot?
Written with AI. OpenAI's descriptions were checked against its published crawler documentation in October 2026. Operators change these names and rules, so check OpenAI's own pages before you rely on them.