I’m testing Perplexity for a research workflow where I need both a readable summary and the underlying results in a form my application can inspect. I can see citations in an answer, but I’m not sure whether the API exposes the search results as distinct data or mainly returns a synthesized response with references.

For anyone who has built with it, how much control do you get over the raw search results and their metadata? I’m especially curious whether they can be stored or processed independently from the generated answer.