But is it deterministic? That has been my main complaint. Every time you see someone else’s results try the same thing and see what you get. Odds are it will be very different and that is a problem.

The best example I can give is searching for hardware documentation for chips. If some asshat company pays google and m$ for promotion it does not matter what I search for I will not get relevant results. I am never searching for a product to buy. Every result will be for this irrelevant shit company and filler junk both engines know to be incorrect. It does not matter where I search. There are effectively only two crawlers that exist. My inquiry is unique enough that all searches fall into the same irrelevant garbage echo chamber. I mostly quit trying to program stuff I’m interested in because of this behavior. I haven’t found a way around it that works. I had the same problems regardless of VPN, or public SearX instances. I don’t know how self hosted would change this. I get the impression that some algorithm jokingly allows some people through once it establishes their identity and a monetization path. When the network and device are never used for purchases, it’s shit all the way down.