What you can conclude
- A ChatGPT-associated fetch requested the recorded page.
- You can inspect the response status and request time.
- Published network evidence can support verification.
Inspect user-triggered OpenAI page retrieval without mixing it into autonomous training or search crawler totals.
ChatGPT-User represents page retrieval associated with an action a ChatGPT user initiated.
This identity is more closely connected to a user-triggered fetch than an autonomous crawl, but the server request still does not reveal the prompt, answer, or whether the fetched content was shown.
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/botTalivia looks for the stable ChatGPT-User token in the normalized User-Agent.
User-agent: ChatGPT-User
Allow: /
# To block instead:
# Disallow: /Use robots.txt to express an allow or block preference for this crawler.
When the integration supplies a source IP from a trusted network boundary, Talivia can compare it with provider-published ranges.
View official sourceTurn on collection in website settings.
Send the minimal event from trusted server code or middleware.
Review pages, statuses, trends, and verification confidence.
Talivia first matches the ChatGPT-User User-Agent token. When a trusted source IP matches provider-published ranges, the request can receive stronger verification confidence.
Add a rule for ChatGPT-User in robots.txt and check the provider’s official documentation for behavior. robots.txt expresses crawl preferences; it is not access control or identity verification.
Talivia stores the hostname, sanitized path, method, User-Agent, optional status, classification, and verification result. It does not store page bodies, cookies, authorization headers, query strings, or raw IP addresses.
Start recording real requests with server-side tracking.
Add your website