If I block GPTBot, will my brand disappear from ChatGPT?

No — and this is the most common confusion about the bot. GPTBot only gathers training data. ChatGPT Search citations run through OAI-SearchBot and Bing, and live page reads run through ChatGPT-User. You can block training and stay fully searchable.

How do I verify a GPTBot visit is genuine?

Check the requesting IP against the ranges OpenAI publishes in machine-readable form on its site. Spoofed GPTBot strings are common in scraper traffic, so never allowlist on user-agent alone.

What Is GPTBot?

Last updated: July 5, 2026 · Menra Team

GPTBot is OpenAI's web crawler for collecting model training data. Announced in August 2023, it identifies itself with a user-agent string containing GPTBot and a link to OpenAI's documentation, respects robots.txt directives, and crawls public pages to improve future GPT models — not to power live search answers.

What GPTBot does and does not control

GPTBot sits in exactly one lane of OpenAI's three-bot system. It feeds training corpora; it has no role in whether ChatGPT Search cites you (OAI-SearchBot's job) or whether ChatGPT can open your page during a conversation (ChatGPT-User's job). OpenAI documents all three bots separately precisely so publishers can make independent choices. The practical upshot: a robots.txt decision about GPTBot is a decision about your presence in future model memory, on a timeline of training cycles, not about today's answers.

The allow-or-block tradeoff

Allowing means your content can enter training data, which supports long-term parametric familiarity — models describing your brand accurately without retrieval. For most brands seeking AI visibility, that compounding recognition is worth more than the content's training value is worth protecting.
Blocking is a legitimate IP position for paywalled or licensing-minded publishers, and adoption was rapid: within a year of launch, tracking projects found on the order of a quarter to a third of the top 1,000 websites disallowing GPTBot.
Blocking by accident is the failure mode to avoid — blanket AI-bot rules copied from blocklists have cut brands out of training data they actively wanted to be in.

Technical profile

GPTBot fetches raw HTML without executing JavaScript, honors Disallow and Allow path rules under the GPTBot user-agent token, and OpenAI publishes its egress IP ranges for verification. Crawl frequency varies with site prominence; most sites see modest, well-behaved request rates.

Example

A SaaS company wanting maximum ChatGPT visibility sets User-agent: GPTBot / Allow: / for its marketing and docs while disallowing /app/ paths. Training crawls get the public story; private surfaces stay out. Sibling bots are profiled in the glossary.

Frequently asked questions

If I block GPTBot, will my brand disappear from ChatGPT?: No — and this is the most common confusion about the bot. GPTBot only gathers training data. ChatGPT Search citations run through OAI-SearchBot and Bing, and live page reads run through ChatGPT-User. You can block training and stay fully searchable.
How do I verify a GPTBot visit is genuine?: Check the requesting IP against the ranges OpenAI publishes in machine-readable form on its site. Spoofed GPTBot strings are common in scraper traffic, so never allowlist on user-agent alone.

Keep exploring

See how AI engines talk about your brand — track mentions across ChatGPT, Perplexity, Claude, Gemini and 5 more. Start with Menra