this looks VERY interesting!
How does this show customization of almost anything? Looks like you can only apply one optic at a time and they are very limited in scope. Like I’ve never cared about hacker news. I don’t care to remove the 10k most popular sites, in fact it’s want the opposite. Additionally I’d want to mix and match optics. Also let optics be customable. Like why 10k top sites. Why not top 1k or bottom 3000k. How many are there even? The number means nothing if you don’t understand what you’re left with.
You could create your own custom or multi-optic by looking at the existing ones, merging them in a text editor, and hosting the file somewhere accessible, like github.
See https://trystract.com/settings/optics , the top 10k list also has a comment telling you how it was generated, so you could just create your own top 1k or bottom 3K list, just change the
DiscardtoBoost(0)if you want to limit your results to only these sites.I really think having to merge random optics together and then setting them when the system could just allow them to merge them or have variables in them is far too much work. The system should be created with customization for the user, not the filter creator.
Kagi does exactly what you’re describing. It’s what I’ve been using.
Been hearing about it a lot, so your comment made me check them out.
First 2 links I clicked on on their homepage are 404s. Like, what. I don’t have the words for what kind of impression that gives.
Specially for a paid search service, hard pass that they can’t even point their links correctly.
Hmm I just checked, they’re all live and their status page for each link has no outages. I would check any content blockers etc. that you have, I suspect it’s a problem on your end. They do use different domains for their blog, feedback, and help KB etc.
Weird, they do, but they redirect for me and the final URL is different than what you pasted.
https://help.kagi.com/kagi/getting-started/
https://help.kagi.com/kagi/why-kagi/why-pay-for-search.html
My best guess is that a DNS record is messed up on their end, and since I’ve been to those pages before relatively recently, the cname or A-record is still cached for me.
404 for me too
I am not the developer of this project.
Lol it searches through stack, github, fediverse. Really nice.
I tried finding information on what indexer they are using. Are they using their own?
Edit: says this in the readme:
The commoncrawl organization for crawling the web and making the dataset readily available. Even though we have our own crawler now, commoncrawl has been a huge help in the early stages of development.
