Allow ChatGPT search without automatically allowing model training.
OpenAI documents separate crawlers for search, model training, and ad validation. Build an explicit robots.txt policy for each use instead of copying a blanket AI-bot block.
Choose each use independently.
The default policy allows ChatGPT search discovery while blocking model-training and advertising crawlers. Change it to match your actual publishing policy.
# OpenAI search visibility
User-agent: OAI-SearchBot
Allow: /
# OpenAI model training
User-agent: GPTBot
Disallow: /
# OpenAI ad landing-page validation
User-agent: OAI-AdsBot
Disallow: /
What does your live site allow now?
Fetch the public /robots.txt file and resolve the homepage rule for each OpenAI crawler. This checks crawler instructions, not firewall access.
What each rule controls
Crawler purposes and current user-agent guidance: OpenAI's official crawler documentation.
Installation checklist
- 1
Copy the generated block and merge it into the existing file at
https://yourdomain.com/robots.txt. Do not erase unrelated rules. - 2
Open the public file and confirm the exact rules you intended are present in the server response.
- 3
Check your CDN, firewall, and bot-management settings. A permissive robots file cannot override a network-level block.
- 4
Keep private or paid material behind authentication. The Robots Exclusion Protocol is a crawler preference, not access control.
- 5
Recheck after CMS, security-plugin, or CDN changes; those systems can generate or replace crawler rules.
Protocol semantics and security limits: IETF Robots Exclusion Protocol (RFC 9309).
What this cannot guarantee
Allowing a crawler makes access possible; it does not guarantee crawling, indexing, citation, recommendation, ranking, or traffic. Blocking a named crawler also does not protect content from every bot or user-directed request. Test the public response and your infrastructure policy together.
Need other providers too? Use the multi-provider AI crawler checker for OpenAI, Anthropic, Perplexity, and Google policy tokens.