But it returns the same results google does. I even asked people on here to give me the queries that kagi is best at. Even then google still returned the same results.
Like the OP said, Google's shittiness is a choice and one they can turn off whenever too many people look like they are about to leave.
Of course the act of visiting Google is the beginning of the monetization. It’s not just an adware company it’s a spyware company as well. This is why their service is free.
And why Kagi is not.
- https://guix.gnu.org/manual/en/html_node/Web-Services.html - https://packages.guix.gnu.org/packages/nginx/ - https://guix.gnu.org/manual/en/html_node/Certificate-Service... - https://guix.gnu.org/cookbook/en/html_node/Setting-up-NGINX-...
And some results from the GNU Guix mailing list. !guix searches with a custom lens[0] that I added for Guix specific documentation and resources like the mailing list archives, and the IRC log archives.
I can combine this bang with the quick summary feature (accessed either by adding !q to the search as well, writing a query that ends in a ?, or pressing the q button after loading the results), which is essentially the same as the info boxes that Google will add sometimes, except it only shows up on Kagi if you specifically request for it.
I have also mapped versioned documentation for some things to the latest documentation using url redirects[1]. I have also manually hidden or raised lots of sites that I don't care about or care a lot about. This manual curation makes it so your search results get better over time, without having something like Google try to deeply understand who you are and what you want to try to figure out what search results are best for you.
Kagi specific features have been very integrated into my workflow personally. I definitely find the subscription worth it.
[0] https://kagi.com/lenses/l7mPOuJp7zljHquBjsekFn6dM9Thw1A8