Some people might have issue with their push into LLM AIs with their Assistant. I personally don’t care for it and am happy that by not using it I’m not subsidising other peoples use of the paid APIs they use. But I’ve seen some people take issue with the development time being taken up by it at all.
I’m a newish Kagi user and I find myself using the LLM about as frequently as search itself.
Sometimes I search for things I know I am looking for. Other times I don’t know quite what I’m looking for or I know in advance that I’m not likely to find it—so I chuck it at Llama 4 Maverick and it usually gives me something useful.
I had no plans to use the LLMs until they opened it up on my tier. At this point however, it’s half the value I get out of Kagi.
I am a proud user of their assistent. It provides access to all models that I am interested in (basically only Sonnet/Opus) with stronger privacy guarantees than many of their competitors. Their UI/UX has definitely room for improvement. However, I find it pretty useful.
It’s fair to be concerned, but web search is probably the most likely field to be completely disrupted by AI. I mean if I’m asking a question, and would find an answer in the first page of search results, it’s pretty likely AI surface the same information more quickly than me sifting through pages of search results.
So I think it’s fair for them to at least have people thinking about that. Plus, the features they have aren’t intrusive and are completely optional. Like, it’d be dumb for a company to not spend development time on a threat that has a decent potential to shrink their (already captured) market.
Even if search gets me the right results at the top of the stack, I have to click into them. Something that can (correctly) summarize all of the results with references is absolutely an improvement.
I am one of those worried people (and have voiced that opinion here on HN before).
However they have been on that course for a good part of the last year though, and they ultimately deliver a good product, which is what matters at the end of the day. Kagi mostly feels "feature-complete", and I'd rather have them spend time on the AI projects than trying to be too inventive and overloading the core product, which is the route many other startups take when they get to that point.
I don't think that the core product is nearly complete, even though I'm happy to keep paying for it. I still find myself falling back to Google for things like weather, flight info, or restaurants. I also put a family member on it that has to search for home products regularly as part of their contractor job. Google's decline was making that impossible. They wanted kagi to surface interesting results beyond SEO'd junk sites, which it mostly failed to do (though still better than Google).
