Please support regexp or directory-based blocks/raises/lowers

warpspin

I would love to be able to not only block/raise/lower domain wide but also, maybe via regexpes, on specific paths. As an example, I often have to lookup APIs on ag-grid.com - the documentation there exists in multiple variants, depending on the JavaScript framework in use, in different subdirectories on the server

I'd love to be able to exclude the variants I do not use, but they all of course share the same domain name. Regexp-based blocking/raising would be great but I'm aware this probably has a high performance impact, so if regexp is not possible, maybe at least something based on subdirectories could be done.

Pretty sure other good use cases for such a feature would exist.

Vlad

warpspin One thing to consider in the meantime is to create a lens that will have sites (and paths) that work for your use case.

warpspin

Vlad Thanks, but it seems lenses does not allow me to pick specific subdirectories on a domain, too?

Vlad

warpspin They do, you can do something like site.com/path

jan

Reddit (and other user content sites like it) is an incredibly useful tool. But the content and quality of subreddits is wildly different from one to the next.

It therefore makes no sense to block the entire domain "reddit.com" because some parts of it I definitely wish to see in my results.

However, I wish to block certain subreddits from showing up in my personal results.

The ability to block/downvote/normal/upvote/pin subreddit-by-subreddit would make this site and Kagi a much more safe and usefull tool.

The tech already exists: it's the same now for separate domain names.

warpspin

That's basically a special case of https://kagifeedback.org/d/1845-please-support-regexp-or-directory-based-blocksraiseslowers :-P

Kraktoos

I am from Portugal and us, portuguese people, have this issue where most of our searches end up in Brasilian websites which is not ideal (much of the content searched for is different, as well as is grammar and, to be honest, there are a lot more of low quality Brasilian websites) so I would like to, instead of having to de-rank every domain that ends with .com.br that I find, having something like regex as an option to allow for things like this.
I would also guess that if I were to de-rank all .com.br websites but then specify a domain and raise it, the more specific rule would apply so it would rise (this might make it harder to implement depending on how it has been done now).

Ex: raising a whole TLD or a domain that contains specific words in it.