Does ChatGPT-User respect robots.txt?

OpenAI distinguishes user-initiated agents from crawl bots; applicability can be partial or different from GPTBot. Treat the official bot docs as authoritative and consider additional access controls for sensitive paths.

Is ChatGPT-User used for training?

ChatGPT-User is associated with live browsing during conversations, not bulk training crawl like GPTBot. Training-oriented collection is primarily discussed under other OpenAI agents—confirm on the official bot page.

Will blocking ChatGPT-User stop ChatGPT from knowing my brand?

Blocking stops or reduces live page fetches under that agent; it does not erase prior training data or third-party mentions. Brand knowledge also comes from public citations, reviews, and other crawlers.

Does ChatGPT-User affect Google SEO?

No direct effect on Google rankings. It is an OpenAI user-agent, not Googlebot.

How often does ChatGPT-User hit my site?

Typically when users or product flows trigger a fetch—often lower and spikier volume than training crawlers. Check access logs for the ChatGPT-User token to measure real traffic.

Should I allow ChatGPT-User if I block GPTBot?

That split is common: limit training crawl while still allowing conversational fetches of public pages. Implement two separate robots.txt groups so the policies do not collapse into one.

ChatGPT-User

OpenAI · Fetches pages when users browse the web during ChatGPT conversations

By Vinespire Editorial Team, Editorial · Last reviewed July 18, 2026

Update: Verified against OpenAI operator documentation for ChatGPT-User.

See our sourcing methodology →

This page is the operator/profile guide for ChatGPT-User. To paste a full robots.txt and check ChatGPT-User plus other AI crawlers in one pass, use the Robots.txt AI Crawler Validator. Short definitional entries also live in the AI search glossary.

Related free utilities are in free AI SEO tools.

What is ChatGPT-User?

ChatGPT-User is OpenAI’s user-agent for live browsing: when a person asks ChatGPT to look at a URL or the product retrieves a page in the context of a conversation, fetches may identify as ChatGPT-User rather than GPTBot. That makes it a live-browsing agent, not a bulk training crawler. OpenAI’s bot documentation distinguishes training/crawl agents from user-initiated fetches; operators should read the current docs carefully because robots.txt applicability can differ by agent. If you care about how ChatGPT experiences your site when a user explicitly opens or cites a page, ChatGPT-User is the relevant token. It should not be confused with OAI-SearchBot, which is aimed at search-style indexing for ChatGPT search features.

How to block ChatGPT-User

Add this group to robots.txt to disallow ChatGPT-User. Path rules can be narrowed if you only need to protect parts of the site.

User-agent: ChatGPT-User
Disallow: /

How to allow ChatGPT-User

For most public marketing and documentation sites, allowing well-behaved AI agents is the default recommended stance for AI visibility—while remaining a factual robots.txt Allow rule you can reverse later.

User-agent: ChatGPT-User
Allow: /

Should you block ChatGPT-User?

Live-browsing agents are usually lower volume than training crawlers and are tied to real user actions. Many sites allow ChatGPT-User so conversational tools can retrieve accurate page content when asked. Blocking may frustrate users who ask ChatGPT to open your docs or product pages. OpenAI’s documentation has historically drawn finer lines for user-agent robots.txt behavior than for GPTBot—verify the current official page before treating a block as absolute. For most public content sites, allowing ChatGPT-User is low risk; block only with a clear policy reason and test the user experience.

Category: Live browsing

Official documentation (OpenAI)

Last verified: 2026-07-18. Re-check operator docs after major crawler announcements.

ChatGPT-User FAQ

OpenAI distinguishes user-initiated agents from crawl bots; applicability can be partial or different from GPTBot. Treat the official bot docs as authoritative and consider additional access controls for sensitive paths.