Found several junk search results in a search query

lacikawiz

Steps to reproduce:
Do the following search:
https://kagi.com/search?q=google+maps+not+working+in+new+chrome&dr=2

Expected behavior:
To get good, spam free search results.

However, there are several results that have the following strange things:

The title is grammatically incorrect, words are in wrong sequence (eg: "Windows Google Working 10 Not Chrome")
The domain is made up of four pieces (eg: eyr.leggings.an.it ), always ending in .it
When the link is clicked, it does a very brief splash screen and then tries to redirect some really strange address (see the second screenshot) which is caught by my adblocking plugin

These kinds of search result should be blocked entirely because they are nothing but spam.

Debug info:
Google Chrome Version 103.0.5060.134 (Official Build) (64-bit)
Linux Mint 20.3 Cinnamon
EUROPE-WEST2

Image/Video:

Vlad

lacikawiz Seems like recent phenomenon, related to this discussion https://news.ycombinator.com/item?id=32201420

Kagi does not crawl the entire web so it is not possible for us to tell if it is a spam domain, but we can try looking for a blacklist.

seventhwave

Thinking about that problem, it seems to me that the only feasible way for Kagi would be to integrate with other sources of truth, like Vlad said: blacklists, or something similar (but imho common blacklist will be fine, at least in the beginning).
About the "selective domain block filter" as suggested by mesaoptimizer, I've to admit that I'm quite biased: I'm from Italy, and .it is my "national" domain, and my results are usually full of ".it" domains.
So, I'm quite sure that no italian Kagi user will never filter by ".it" domain trying to catch the spam; and the same will be for Kagi french user and ".fr" domains, or australian users and ".com.au", etc...
So my proposal will be: integrate blacklists, if it's possible for Kagi architecture.

mesaoptimizer

I think you are right, https://www.register.it/ is offering free domains for 1 year since some time.

-- https://news.ycombinator.com/item?id=32201996

Perhaps a really powerful filter would be one where one can block certain domain suffixes (like this ".it" one) entirely from appearing in search results. This filter doesn't need to be a Kagi default, but as a user willing to sacrifice a few false positive ".it" domains to remove this entire block of domains out of my search results, I would appreciate this feature.

Vlad

mesaoptimizer Can you make a separte suggestion for advanced blocking rules such as *.it