As an aside, I tried to love ddg before using kagi but I'd always revert to google with !g and end up not finding anything because google has sucked for the past 5 years

I've been using kagi and I've noticed that the only time I ever switch to google is for things that are truly local to where I am at the moment (restaurants, etc..).

Kagi is really great.

I've also been testing perplexity and I'm not that convinced yet. The only thing it's been decent at is finding relevant studies.

This is very similar to my experiments, and something I was mulling over enough when I wrote the original comment, that I wrote a rant about it:

https://pdx.su/blog/2024-12-29-using-kagi-for-a-bit-over-a-y...

The thing I don’t like about Kagi is how you have to be logged in with an email to search.

I wish they’d setup something like Mullvad where you get an account number, pay for said account, and that’s it.

What's the difference? It's still a permanent ID tied to your account and linked to your payment. Just generate a random email address if that matters.

Kagi and mullvad make the same claims to privacy and non-tracking of accounts. What's the difference between a random numeric and a random alphanumeric account identifier?

Edit: actually I just remembered that my android devices are linked to my Kagi account with a token. Kagi gives you an alphanumeric token you can use in your query URL to authenticate searches. That's how my homescreen search widgets are set up, no email required.

Last I looked, it's not a verified email, you can use it as a random id, and they even documented that. (Maybe it changed?)

They have paid search, too, which might filter (or mark) the garbage. It required knowing how more of their stack works than I wanted to learn. I couldn’t find any docs to fix the problem I had. So, I ended up using SerpStack (IIRC) with custom rendering which worked well.

That said, is anyone using the Google Search API to try to cut out a lot of the junk the author is talking about? How well has that worked?