Are you scraping ChatGPT?
No. ChatGPT and Perplexity are queried through their own official APIs. The two Google surfaces are not, because Google publishes no API for them — and we say so on every edition rather than blur it.
ChatGPT is asked through the OpenAI API. Perplexity is asked through the Perplexity API. Both are the vendors’ own interfaces, used as intended, and both return the answer together with the sources it was built from — which is where the citation data comes from.
Google AI Overviews and Google AI Mode are different, and this is the part most measurements leave vague. Google offers no API for either surface. They reach us through a third-party SERP-data provider, and that provider’s own documentation describes what it does as scraping. We do not describe it as anything else.
That distinction is printed in the method block of every edition, engine by engine, alongside the model each one ran on. It is there so you can check it rather than take our word, and so a technical reviewer at your company finds it from us before they find it somewhere less comfortable.
Why it matters to you
An answer sourced through an official API and an answer sourced through a scraper are not equally durable. The Google surfaces could change how they render tomorrow; the API surfaces are contractual. The method page names which is which for every engine in every edition, so a figure that depends on the fragile route is identifiable as such.