Similarly, I had no idea about this feature, but I now know I want it... I think I just need it to be cheaper, maybe $5 unlimited (or enough for me, which I think is certainly higher than 300, probably higher than 1000, at least without behaviour change) and I'd be in
I assume it's a bit chicken and egg - marginal costs of new customers/extra searches is relatively low, it's the crawling & development/SRE that's expensive - so should (or at least could) come down as more people adopt it, driving further adoption from relative cheapskates like me?
It seems great, but I use DDG 'bangs' (which are also a Kagi feature) loads of times a day, rather than going directly to websites - especially !w & !wikt, !arch & !aur & !archpackages only slightly less - I'd blow through the monthly limit in no time on that alone. (Which also seems a bit unfair, they're just redirects, not using index, there's an argument they shouldn't be included? I suppose if you're so motivated you can work around by using Firefox's feature for it instead of search engine anyway.)
Edit: oh wait, I just found in the FAQ that 'bangs' actually aren't counted as searches. I may have to trial it, see how many searches I actually use how quickly. (And how reasonable the 1.5ç overage seems in my usage.)
I was worried about the 1000 search limit, and it has so far (since beta) proven not to be a problem. I use Kagi for all my searches on all machines and my phone and average 800 or so a month. I haven’t modified my behavior at all, and consider myself a heavy search engine user.
I have a theory that I would actually need to do twice as many searches with Google or DuckDuckGo or whatever, since the SEO spam would force me to do more term refinement. With far less of that (and a tiny tiny bit of settings tuning) I get better results much quicker. I’d test it but I have a job to do and can’t swallow the idea of going back to how bad the other viable options really are.
That sounds appealing, but there are so many $10 and $15 monthly warts on my balance that the burden of adding another (with even a minor overcharge fee) feels pretty high. I wish their lower tiers had double the current limits.
> I have a theory that I would actually need to do twice as many searches with Google or DuckDuckGo or whatever
FWIW, I counted searches I made with DDG (by parsing my FF history export) before joining the beta, and it was slightly over 1k, with Kagi my searches are in the 700-800 range.
> I have a theory that I would actually need to do twice as many searches with Google or DuckDuckGo or whatever
This is a great point that should really be highlighted more in their marketing. It was obvious once I read it and shows two obvious benefits: 1. There is probably zero issue with fitting in to the 1000 searches per month. 2. It saves time for many of the searches we’re already doing.
> 1. There is probably zero issue with fitting in to the 1000 searches per month.
You are never limited to 1000 monthly searches, this is a great misunderstanding about Kagi. After passing your threshold, you are charged 1.5c per query.
I think that’s clear enough on the details of the plans. What I meant is fitting in without incurring extra charges (no point in arguing that 1.5c is small, that’s not the point)
I’m a former DDG user and current paying Kagi customer. One critical difference I’ve noticed is that frequently used bangs with DDG, but almost never with Kagi.
I think this is because over a very short period of usage, I applied block/lower/raise/pin to various domains, and now my search results are almost always perfect.
That said, I understand the value of going directly to a site. On that front, Kagi makes it trivial to define new bangs.
Lately I’ve moved to doing this at the browser level instead of via a search engine. It’s not a huge speed difference, but in principle it seems wrong to have an extra round-trip out to a search engine to redirect me to another site, when I can set up the keywords locally and go directly. Firefox makes it easy to do: https://support.mozilla.org/en-US/kb/how-search-from-address...
I've used DDG for years before switching over to Kagi and I was accustomed to having to use the !s bang daily (and !g before I found out about !s). On Kagi I literally never use it, since it already incorporates Google search results and filters out a lot of crap.
I've also defined some custom bangs to search private Jira and Confluence instances with some default filters applied, because the default search on those instances was driving me insane. It took me a long time to convince myself to pay $10 a month for a search engine, but I've never looked back.
I'm currently experimenting with hostname rewrite rules.
I'm at the $10/mo level, and am totally happy to keep paying. I rely on search throughout the day for my job, and for my personal interests. $10/mo for an ad-free tunable search that provides good results is a completely worth it. And, I'm more than happy to vote with my money to support the web that I want to exist.
Otoh (for contrast) I run adblock and sponsorblock on youtube, and if they start blocking traffic from people like me I'm going to cut back my YT viewership significantly (and maybe try harder to look for stuff on nebula instead, which I never browse). While I do get some educational value from YT, it's limited, and I guess I just fundamentally object to the ad-first business model.
You should install the Kagi browser extension and try out the Summarizer functionality on YouTube videos. I’ve found it pretty useful for videos that are needlessly filled with fluff.
Oh really? I just commented about how this misconception made me change my behavior. Wish that was more clear, or that I had higher reading comprehension.
In general, assume that we pass down the savings to the user whenever we can.
Redirecting a bang costs us basically nothing, so it does not count as a search.
Also reloading the same search within a short time (~2 minutes, for example coming back in browser after clicking through a search result) does not count as an additional search as we served cached results.
Another tip: We have decent documentation (that is also open source and editable) and you can access it quickly from Kagi with the !help bang, for example
> In general, assume that we pass down the savings to the user whenever we can.
This is great to hear but it doesn’t really address the point you’re replying to:
> Wish that was more clear
As someone who is just seriously noticing Kagi because of this HN post, I have no brand impression that tells me you’re trying to pass on the savings (most companies don’t) and the marketing material even makes me think you’re targeting a premium price point with healthy margins.
I’m also going to suggest a do not do: Do not make cost savings a public-facing core value. Racing to the bottom usually erodes the value prop while simultaneously opening the door for low-quality competitors.
I would suggest (and caveat: I’m not a marketer, just someone who has seen a lot of brands come and go) either of these directions:
1. Be clear about the cost savings when talking about the specific features. Example: Add a new heading to kagi/features/bangs.html that says “Bangs are Free” and talks about how bangs do not count towards your searches
2. Remove it from being part of your public values altogether. Lean in to the value prop of having a cost and let users self-discover what’s free. If it’s the right choice, highlight it privately: make it a line item on the monthly invoice with something that describes why they are free or put it in onboarding material.
I didn't say it was a 'killer feature', I said 'now I know I want it'. There are very many things I want but not at the price they cost!
I don't think coffee's a great analogy - it has far more utility for me, I don't buy it as a service, and it costs me roughly £24 (what, $30ish) a month in beans.
I'm not seriously suggesting it because I think comparing prices of different things to determine value very quickly gets silly, but a closer comparison might be Netflix: Kagi would cost me about as much per month for search as Netflix. Silly, as I said, but if I tried really strictly to pay for things according to relative value or utility to me, there's no way that would make sense.
Part of the problem is asking me to switch from free I suppose - if DDG suddenly started charging me $5 I'd be more likely to pay Kagi $10 than I am today if that makes sense.
People always compare with Netflix, but that makes no sense at all to me. Is Kagi comparable to Netflix because they are both presented to you on a digital screen? I see nothing else that makes them comparable.
Laundry detergent and breakfast cereal both come in a box in the supermarket, but it doesn't make the products comparable. A house and a car both have a door and you usually sit inside, but it doesn't make the products comparable.
Netflix is a great bargain (for those who enjoy their offerings), and I think people are shooting themselves in the foot by dismissing other great services because they've somehow tricked themselves to thinking they should be compared to Netflix only because they're offered through a digital screen.
I have gone over the limit (on the grandfathered 1500 search plan), and paid the extra $1-2 for it. It's fine.
The cost of search has made me think about what terms I use now and what I search for. Google made me lazy, I now use my phones world clock app rather that searching for "current time X", and word my searches more thoughtfully.
Defaulting to DuckDuckGo and using a bang for kagi when I don’t immediately get what I want would be nice. Unfortunately there isn’t one. I guess I could use an extension that lets me add personal ones.
I subscribe to the cheapest plan, and only use Kagi for non-trivial searches. That’s more than enough, I don’t need a paid search engine when I just type a website’s name instead of its url.
I set it up with DDG/Google as default, and for Kagi I start with “k “.
I assume it's a bit chicken and egg - marginal costs of new customers/extra searches is relatively low, it's the crawling & development/SRE that's expensive - so should (or at least could) come down as more people adopt it, driving further adoption from relative cheapskates like me?
It seems great, but I use DDG 'bangs' (which are also a Kagi feature) loads of times a day, rather than going directly to websites - especially !w & !wikt, !arch & !aur & !archpackages only slightly less - I'd blow through the monthly limit in no time on that alone. (Which also seems a bit unfair, they're just redirects, not using index, there's an argument they shouldn't be included? I suppose if you're so motivated you can work around by using Firefox's feature for it instead of search engine anyway.)
Edit: oh wait, I just found in the FAQ that 'bangs' actually aren't counted as searches. I may have to trial it, see how many searches I actually use how quickly. (And how reasonable the 1.5ç overage seems in my usage.)