Skip to content
Answer Rankings Answer Rankings Answer Rankings

GPTBot: What It Is and Whether to Block It (2026)

GPTBot is OpenAI’s primary web crawler. It gathers content used to train and improve the models behind ChatGPT. If GPTBot has never crawled your site, ChatGPT’s models have little first-hand knowledge of you — and blocking it in robots.txt keeps it that way.

User agent

Section titled “User agent”
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; GPTBot/1.2; +https://openai.com/gptbot

Robots.txt token: GPTBot. OpenAI publishes its crawl IP ranges, so a claimed GPTBot hit can be verified against them — useful because fake “GPTBot” scrapers exist.

OpenAI’s three crawlers — don’t confuse them

Section titled “OpenAI’s three crawlers — don’t confuse them”
BotJobRespects robots.txt
GPTBotModel training corpusYes
OAI-SearchBotChatGPT Search index + citationsYes
ChatGPT-UserLive fetches when a user asks ChatGPT to open a pageYes (fetches on user request)

Blocking GPTBot but allowing OAI-SearchBot means ChatGPT can still cite you in search results but the underlying model won’t “know” you. Most businesses want both allowed.

Allow or block?

Section titled “Allow or block?”
# Allow (recommended for visibility)
User-agent: GPTBot
Allow: /


# Block (content protection)
User-agent: GPTBot
Disallow: /

Block only if keeping content out of training data matters more to you than being present in ChatGPT’s answers — a real trade-off for paywalled publishers, a costly own-goal for almost everyone selling products or services. Many sites added a blanket block in 2023 and forgot; it is the single most common failure our AI visibility checker finds.

Section titled “Related”