When I ask the same question in Perplexity more than once, I sometimes get a different set of sources or a different angle in the answer. I’m not sure whether that’s because the web has changed, the query is interpreted differently, or the search itself is meant to vary.

For research where I need to compare results later, is there a way to make a Perplexity search more repeatable across separate searches? I’m interested in what other people have observed, not just whether the wording of the prompt matters.