The author did not mention that you as a end user can customize rankings for every website and build custom searches (lenses), which are in my opinion the features that make Kagi unique and more useful than other meta search engines. And to my knowledge you cannot replicate those in SearNXG.
You can somewhat exclude websites from results if you self host SearXNG, but only by modifying configuration files… Which is a bit annoying.
one detail that’s often glossed over is that Kagi primarily functions as a middleman between you and other search engines, like Google. In other words, it’s what we call a metasearch engine.
Yeah, I’ve known this. I’m… ajdlfjk about it.
What it can do, however, is make content less visible or harder to find, based on the criteria it uses to filter and rank results from its sources.
Aaaand this is kinda the secret sauce. While the starting material is what you get in other places, Kagi allows you to improve it with your own rankings and filters on top. For example, every time I find an AI slop review website, I immediately downrank it in my results. I also uprank trusted sites. This at least tilts the scales in my favor when searching, instead of accepting what Google wants me to see.
But, yeah, I agree Kagi is not discovering new material not known to Google, but it does have a higher chance of surfacing it.
it’s not fundamentally different from what other search providers are already doing.
I haven’t compared search engine features in a bit. Maybe it’s time.
I encourage you to try SearXNG
OK, I will.
I tried Kagi (100 query free trial if you supply a fake email address, I didn’t test whether a real one works). It was nice in some ways but gave about the same results as Duckduckgo. I didn’t subscribe.
No fucking way am I clicking that URL.
Others have explained the unicode-in-URL aspect sufficiently, but I can speak to the author of the site somewhat. His or her blog posts have hit the fediverse several times before. They’re often insightful and skeptical, highly privacy conscious. I hope they don’t mind if I take this part from their FAQ:
Can I trust the information on this website?
No. And you should never trust any single website or entity. Especially not the ones that have sponsored content or have no academic/professional background in the topics they post about.
Take this information as mere pointers into different directions, that scratch the surface and ultimately provoke your itch to find out more about the individual topics. Do your own research and come to your own conclusions.
It’s punycode, it displays as japanese characters in a real browser.
https://en.m.wikipedia.org/wiki/Punycode
Thx TIL
I didn’t know about that, thanks for sharing.
For those curious, the characters are katakana (the syllabary often used in Japan for foreign words, onomatopoeia, etc) and they’d be read as “ma-ri-u-su”, which is possibly intended to represent “Marius” under Japanese spelling conventions.
in a real browser
Which is a major security risk and you should avoid those “real browsers”.
by displaying Unicode characters an attacker can send you a link that clearly shows its yahoo.com and you see in the browser url that its yahoo.com but in reality its unicode letters that look similar to latin one.
that’s really bad
Not everyone uses English. Many languages can’t be written with only ASCII characters.
xn–gck…whatever is famous in some circles!
So, you monolingual? 🙂
Woah are you a monophobe or something?
Its fine. Trust me /s
Actually probably good on you to not trust everything on the internet.
This individual has a strange url but generally has some really insightful articles.
It’s not really a “strange url” it’s just the way we decided to implement unicode into URLs, which for the purposes of phishing links is catastrophically dangerous, so it’s never really been widely adopted and most sites choose by default to show the raw URL, for safety. This is why we can’t have nice things (and instead need to have xn-dfg344jlb5jsdfl543sdfsd.com)
