I want to start by inviting everyone who are looking into this post to check the git commit linked before reading and after that read their current page, it will make more sense.
I think most people don’t understand the fact that Kagi is meta-search engine which literally collect the results of other search engines and display it and add very small amount of results from their index(tecilis).
They are starting to act shady this year(check the attached git commit to see their changes to their documentation page).
People basically pay them to search Yandex and Brave for them.
Before anyone say but privacy, they are closed source and there is a better open source meta search engines which actually respect your privacy and need your donations more than Kagi shareholders.
Open source alternative:
It’s not fair to call Kagi just meta search engine. As you mentioned they do have their own index, it’s just that they are complementary. They even show you how much of the results are unique results from their index. They’re very open about it.
Privacy I can definitely agree on but the actual search result is not better. I think it’s safe to say that if you’re paying for Kagi, you know what Searxng is and can spin up one if you want. I’ve done it myself and it was clearly not better.
And if you think about it, if you self host Searxng, that means all the search query + your IP goes to search providers which kinda kills the point, unless you use VPN or Tor, both of which have their drawbacks.
If you use hosted instance, you have to trust them. Being an opensource project doesn’t mean a lot when you can’t go through the source code that is actually running.
At some point you have to trust someone. I happened to trust Kagi to some extent.