openai.com AI crawler policy

CORE

AI Access Matrix

AI system	Purpose	Observed policy	Provenance
O GPTBotOpenAI	Model training	Partial	`Wildcard`
O OAI-SearchBotOpenAI	AI search	Partial	`Wildcard`
O ChatGPT-UserOpenAI	User-directed retrievalUser-directed	Partial	`Wildcard`
A ClaudeBotAnthropic	Model development	Partial	`Wildcard`
A Claude-SearchBotAnthropic	AI search	Partial	`Wildcard`
A Claude-UserAnthropic	User-directed retrievalUser-directed	Partial	`Wildcard`
G Google-ExtendedGoogle	Gemini training / grounding policyPolicy token	Partial	`Wildcard`
P PerplexityBotPerplexity	AI search	Partial	`Wildcard`
P Perplexity-UserPerplexity	User-directed retrievalUser-directed	Partial	`Wildcard`
 Applebot-ExtendedApple	Model training policyPolicy token	Partial	`Wildcard`

HISTORY

Policy Timeline

No semantic change yet.The first observation is the baseline. Future policy differences will appear here.

TRUST

What this means

Quoana reports the policy a site publicly declares. It does not prove that a provider actually crawled the site, nor does robots.txt act as an access-control mechanism.

Read the method →