Does Blocking GPTBot Remove You From ChatGPT?

Does blocking GPTBot remove your site from ChatGPT? No. Blocking GPTBot alone does not remove your website from ChatGPT search or ChatGPT answers. It tells OpenAI not to use your pages to train its foundation models, and nothing more.

The rule that can remove you from ChatGPT search is a rule blocking OAI-SearchBot. If you copied a blanket “block all AI bots” rule into your robots.txt file, you may have blocked that crawler without realising it.

You can refuse AI training and still allow your pages to be found and cited in ChatGPT. The two choices use separate robots.txt rules.

OpenAI has four crawlers, and they do different jobs

A crawler is a bot that requests pages from your website. OpenAI uses four different crawlers for four different purposes:

• GPTBot collects content that may be used to train OpenAI’s foundation models.

• OAI-SearchBot powers search results and citations in ChatGPT search.

• ChatGPT-User fetches a page when a ChatGPT user asks about it or asks ChatGPT to visit a particular link.

• OAI-AdsBot validates pages submitted as advertisements.

These names can look interchangeable in a robots.txt file, but they are not. Blocking one does not automatically block the other three.

What happens when you block GPTBot?

A rule for GPTBot opts your site out of OpenAI’s training crawl. It does not opt your site out of ChatGPT search.

For example, this rule blocks GPTBot:

User-agent: GPTBot
Disallow: /

It means GPTBot should not crawl any page on your site. It does not tell OAI-SearchBot to stay away. ChatGPT can still discover your pages through its search crawler, include them in search answers and link to them, subject to the usual factors that affect whether a page is selected.

This is the important distinction: training access and search visibility are separate permissions.

What happens when you block OAI-SearchBot?

Blocking OAI-SearchBot is the rule that affects whether your pages can be shown as sources in ChatGPT search.

OpenAI’s own documentation says that sites opting out of OAI-SearchBot “will not be shown in ChatGPT search answers, though can still appear as navigational links”.

In practical terms, your site may still be mentioned if somebody asks ChatGPT to go to your domain directly. But your pages will not be eligible to appear as the normal web sources selected for ChatGPT search answers.

So the answer to “does blocking GPTBot remove you from ChatGPT search?” is no. The answer to “does blocking OAI-SearchBot remove you from ChatGPT search?” is broadly yes, according to OpenAI’s documentation.

The blanket “block all AI bots” rule is the trap

Many site owners added AI crawler blocks in 2023 or 2024, when the simplest advice was to block every bot associated with AI companies. Some copied a list of names. Others added a broad rule that matched every crawler they did not recognise as a search engine.

That can block GPTBot, OAI-SearchBot, ChatGPT-User and OAI-AdsBot together. If OAI-SearchBot is included, your site loses eligibility for ChatGPT search citations.

A broad rule can be difficult to spot because it may not contain the words “ChatGPT” or “OAI-SearchBot”. Look for entries such as:

• User-agent: OAI-SearchBot followed by Disallow: /

• User-agent: * followed by Disallow: / in a setup intended to block all crawlers

• A long list of AI bot names that includes OAI-SearchBot

• Comments such as “block AI”, “stop AI scrapers” or “no generative AI” above a group of rules

The User-agent line is simply the name of the crawler a rule applies to. Disallow: / means that crawler is asked not to access any path on the site. You do not need to be a developer to check those two lines.

You can refuse training and stay eligible for citations

Refusing to let GPTBot crawl your website while allowing OAI-SearchBot is a legitimate position. It lets you say “do not use my pages for training” while still allowing your content to be discovered for ChatGPT search answers.

These are separate robots.txt rules, and the blank line between each group matters:

User-agent: GPTBot
Disallow: /

User-agent: OAI-SearchBot
Allow: /

If you also want pages fetched when a user asks ChatGPT about them, add ChatGPT-User. And if your business submits pages as OpenAI advertisements, you can allow the advertising validator too. In one complete set:

User-agent: GPTBot
Disallow: /

User-agent: OAI-SearchBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: OAI-AdsBot
Allow: /

Before adding this, remove or update older entries that separately block OAI-SearchBot. A new allow rule will not help if your file still contains a conflicting rule, or if your website platform generates another block later.

ChatGPT-User is slightly different

ChatGPT-User is not the same as the search crawler. It is used when a user initiates an action, such as asking ChatGPT to read a particular page.

Because these actions are initiated by a user, robots.txt rules may not apply.

That is OpenAI’s stated position, not a settled guarantee that applies identically to every request, platform or future product. Treat ChatGPT-User as a separate case from search eligibility. Allowing it does not replace allowing OAI-SearchBot if you want your pages considered for ChatGPT search answers.

How common are these blocks?

Blocking AI crawlers became common enough that old robots.txt files deserve another look — but published rates vary enormously depending on which sites are sampled, so treat any single figure with care.

Originality.AI tracked the top 1,000 websites after GPTBot launched in August 2023: around 5% blocked it at launch, rising to 35.7% by August 2024. A broader July 2026 study by HasData, covering 10,894 domains from the Tranco top 10,000 plus news publishers, found only 7.9% of the general web blocking GPTBot — but 50.5% of news publishers. A smaller June 2026 sample of 107 prominent sites by US Tech Automations put GPTBot blocking at 30.8%.

Those numbers are not contradictory. They measure different webs. The more prominent and publisher-heavy the sample, the higher the blocking rate.

Publishers are the extreme end, and their behaviour is the most relevant to this post. BuzzStream’s April 2026 analysis of 100 top news sites in the US and UK found 79% block at least one AI training bot — but also that 71% block at least one AI live search or retrieval bot, and 67% block PerplexityBot specifically. Those retrieval blocks are the ones that cost citations, and it is not always clear the two decisions were made separately.

News publishers have stronger reasons than most to control how their reporting is collected and reused, so their behaviour should not be read as a picture of every small business website.

Blocking AI crawlers does not hurt your Google rankings

Blocking GPTBot or another OpenAI crawler does not damage your Google rankings. These are different crawlers and different systems.

The cost of blocking OAI-SearchBot is visibility in ChatGPT search and the citations or links that come from it. The cost of blocking GPTBot is opting out of that training crawl. Neither rule is a Google ranking penalty.

Of course, a careless blanket rule that blocks all crawlers can cause wider problems. If your robots.txt file blocks Googlebot, Bingbot or every crawler with User-agent: *, search engines may not be able to crawl your site. That is a separate and much more serious issue, so check that your normal search engine access has not been caught in the same rule.

How to check which case you are in

1. Open your website’s robots.txt file by adding /robots.txt to the domain, such as example.com/robots.txt.

2. Search the page for GPTBot, OAI-SearchBot, ChatGPT-User and OAI-AdsBot.

3. For each name, look at the following Disallow and Allow lines.

4. Check whether an old “block all AI” section includes OAI-SearchBot.

5. Check for a broad User-agent: * rule that blocks the whole site, and make sure your main search crawlers are not blocked.

If GPTBot is blocked but OAI-SearchBot is allowed, you have refused training while remaining eligible for ChatGPT search citations. If OAI-SearchBot is blocked, you have opted out of those search answers, even if GPTBot is also blocked.

One caveat worth knowing: robots.txt is a declaration, not a wall. HasData found that 39.5% of declared GPTBot blocks are not enforced at the server at all. A crawler can also choose to ignore the file, and a site can allow a crawler in robots.txt while a firewall or CDN rule blocks it anyway. Reading the file tells you what you have asked for, which is the right place to start, but it is not proof of what actually happens.

What ours looks like

We take the position this post recommends, and our robots.txt names every answer crawler explicitly rather than leaving them to a wildcard:

User-agent: OAI-SearchBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

Naming them is deliberate. A wildcard already allows them, but writing each one down means a future decision to block AI crawlers has to be made about each crawler on purpose, rather than swept in by a single line — which is exactly the mistake this post is about.

Read your robots.txt today

Most site owners do not know which of these two cases they are in. They remember adding an AI block, but not which crawler names or broad rules it contained.

Go and read your own robots.txt file today. If your aim is to refuse AI training while staying citable in ChatGPT, block GPTBot specifically and leave OAI-SearchBot allowed. Do not rely on a blanket “block all AI bots” rule when the result you want is more precise.

If you would rather not read the file line by line, our free scan reads it as each of these crawlers and tells you which case you are in — no account needed.

For the official crawler descriptions and the latest guidance, see OpenAI’s bot documentation.