I’m a very happy Kagi user, the search results are way better (subjective, I know) for me, that is, I can usually find what I’m looking for in the first few entries.
I’m under the early adopter pricing but I fear that the higher price (and cognitive effort in lower prices to keep track of how many searches I’ve made) might make adoption a bit harder. I really really really don’t want Kagi to go away! I’ve shared with some friends that are also using it and even though they also reacted to the price change, they decided to stay (just like me) because of the early adopters price.
This part is totally unsolicited advice, but I find my self searching quite often for the same thing, like, my browser will remember a search instead of the domain of the thing I’m searching for (i.e. wanna go to namecheap, my browser remembers the namecheap kagi search instead of namecheap.com), maybe there could be some space to optimize repeated searches? Idk, if I’ve made a search X times the just cache it and return me the same result, don’t count that towards the total, and add a little message that the results are cached and if I want to search again click there? I know I could make bookmarks but changing behaviors is a way taller ask than caching. Anyways, best of luck and I really really hope that kagi stays afloat!
Fellow Kagi user here and for me the search results are so similar to Google's that day-to-day I don't notice which search engine I'm using. This is good and bad: Kagi is on par with Google but for me it hasn't enough benefit to pay for it in the long run.
When it comes to your point of repeated searches: I wish my search results would be a combination of global (unpersonalized) ranking and local personalized ranking. The global ranking would filter out spam and useless content but the final ordering of the results could be determined by a locally running algorithm that learns my preferences. For example when searching for Rust it could filter out all gaming related content and only present me the programming language related results.
I’m not the OP, but in my opinion, yes, absolutely. Kagi makes it really easy to completely filter out certain domains, raise or lower the priorities for others, and to even pin results from your favourite sites.
Interesting! I'll have to check it out. Not being able to filter domains automatically filter or prioritize certain domains has become a real point of annoyance for me.
Main one being all these spammy StackExchange mirrors and similar. A lot of ChatGPT-like blogspam that just seem to be pages to get people to go and buy something from Amazon too.
Good point. I use Kagi like it was Google and I'm so used to unrelated spam that it doesn't bother me much (sadly). I cannot tell if it is more or less on Kagi.
I only realized after reading your comment that I can block and boost results in Kagi and I will definitely start using that. I think I did injustice to Kagi in my previous comment, because it does more or less what I wished for without me noticing. (The fact that I pay for it makes it even more embarrassing;-)
Kagi is fairly recent, something like a year old, so having 15 years in tech won't help you here. I doubt they're astroturfing like you're claiming, they actually have a decent number of users and a good following on HN. You can look at the numbers their previous posts on HN did[0].
what if they had a free plan that allowed 100 searches each month. or extend the trial each month if requested. maybe it'd take many users more than a month to get into it and see the value in investing in a paid plan.
We find that 80% of users that end up paying, do that after their first week (35% on their first day!). We are trying to make it very clear that this is not a yet another free search engine to manage both expectations and cost.
> I’m under the early adopter pricing but I fear that the higher price (and cognitive effort in lower prices to keep track of how many searches I’ve made) might make adoption a bit harder.
I pay for Kagi, I had a few minutes of minor outrage when I saw the pricing changes, which in retrospect was childish (for me, others motivations may differ). Iirc, the pricing actually won't change for me given the number of queries I do, and regardless I'll probably just get the highest tier anyway and be done with it. There are always ad supported options for people who want to save money. Personally I prefer to pay what it costs (meaning pay so that kagi can run a viable business) because I know how advertising perverts a business. I hope they can succeed by appealing to people who think this way.
Also re the GP mention of caching, I'm sure I read in a discussion that kagi had tried that and it didn't change the economics.
For me, ~50% of searches are just lazy ways to get to a website I know, like I literally search for hacker news regularly. I'm curious if there's some kind of triage possible, where the lazy site lookups get a simple search that's cost efficient and then fail over to a costly search if they don't find anything. If there isn't something like this ready. Modern search is really just a "portal" with fuzzy matching for most queries, as opposed to genuine "show me a site I don't already know about" and I've never seen that reflected in any discussion about search.
I’m under the early adopter pricing but I fear that the higher price (and cognitive effort in lower prices to keep track of how many searches I’ve made) might make adoption a bit harder. I really really really don’t want Kagi to go away! I’ve shared with some friends that are also using it and even though they also reacted to the price change, they decided to stay (just like me) because of the early adopters price.
This part is totally unsolicited advice, but I find my self searching quite often for the same thing, like, my browser will remember a search instead of the domain of the thing I’m searching for (i.e. wanna go to namecheap, my browser remembers the namecheap kagi search instead of namecheap.com), maybe there could be some space to optimize repeated searches? Idk, if I’ve made a search X times the just cache it and return me the same result, don’t count that towards the total, and add a little message that the results are cached and if I want to search again click there? I know I could make bookmarks but changing behaviors is a way taller ask than caching. Anyways, best of luck and I really really hope that kagi stays afloat!