Hacker News Reader

Comment by bayindirh | Hacker News Reader

Kagi is also building their own index at the background, and mixes these indexes as you search.

When I search Kagi for "Hacker News", results start with this fine text:

    65 relevant results in 1.09s. 47% unique Kagi results.

So, other indexes are fillers for Kagi's own index. They can't target their bots to places, because they don't have the users' search history. They can only organically grow and process what they indexed.

View on HN

How is it possible that a search for "Hacker News" produces only 65 results? There are thousands of pages out there with that exact phrase on it (including many sub-pages of this site).

The first result is almost assuredly the right one, but either they're ruling out a lot of pages as not-what-you-meant, or their index is really small.

It's another feature of Kagi. They know they have thousands of results, but they provide you a single page of most relevant results. If you want to see more because you exhausted the page, there's a "more results" button at the bottom.

Kagi reduces mental load by default, and this is a good thing.

That's an interesting positive spin on what is clearly a cost-cutting feature. I'm on the unlimited search tier, so it hasn't really been a big deal, but it's worth noting that clicking the "more results" button charges your account for an additional search too. Or at least it used to.

I'm also on the unlimited search tier, and honestly never needed to reach for the more results button even once. Kagi always delivered.

Having a low number of results has a net benefit of lower cognitive load for me, so I like how Kagi returns less results, not more.

But "more results" is counted towards a new search, and I didn't know that. Thanks for pointing out.

Reply on news.ycombinator.com