The search engine does indeed seem to be fully custom - the results for niche topics are quite bad, but that at least means it's not a Bing frontend, so it's forgivable. Although I wonder why the authors decided to go 100% custom instead of (partially?) reusing https://github.com/MarginaliaSearch/MarginaliaSearch for some of the core engine.
On an unrelated note: why do some people use LLMs to write all the text in their projects? As an example: https://xonaly.com/how-it-works - it's obvious that the text here is LLM-written, but I genuinely don't understand why. It's one of the pages linked directly from the home page, so it's quite important. At least personally, I always get very suspicious of any project that doesn't even bother to have some human-written content. (I know the whole website clearly looks LLM-authored, but I feel like it's not hard to add some human touch)
Don't get me wrong, I use LLMs myself quite a lot, but I try to always interact with others in my own words. In rare cases where I need to paste some output from an LLM, I always leave a note about it being AI-written.
Oh, and another funny tidbit I just found: https://xonaly.com/is-xonaly-legit/ repeats the same facts a lot. Ctrl+F "Canad", "independent": 10 results, "search engine": 17 results.
The text seems like it's there as a basic low effort placeholder.
Maybe it's still a one person project and they don't want to spend their time crafting reasonable marketing fluff pages, so they relied on an LLM to generate that while they focus on the tech side of things.
This type of stuff just screams, “please do not use me” to me. It seems lazy and full of shortcuts on the surface, even if it might not be in reality. And im very pro AI. I’m just also very anti AI slop and regurgitated outputs.
I wondered if the whole thing was just a vibe coded weekend project. It’s hard to gauge the seriousness and level of effort but without some evidence to the contrary I’m guessing it’s all LLM generated.