OpenAI's reported Aura safety check could block sites from ChatGPT browsing

According to Search Engine Journal, OpenAI runs an automatic site-safety gate called Aura that can stop ChatGPT live fetches, Codex Browser Use and its cloud browser from loading a domain, independently of robots.txt or owner allowlists. OpenAI has not publicly documented Aura's criteria or an appeal process.

• Aura can block sites for ChatGPT, Codex Browser Use and OpenAI's cloud browser
• The decision reportedly works independently of robots.txt and owner allowlists
• No public self-service interface exists to appeal an Aura block
• The report cites a major publisher blocked since May 2026