> We are also thrilled to report that we have achieved profitability. This significant milestone is a testament to our sustainable growth and fiscal responsibility. It demonstrates that our approach of offering a premium, ad-free search experience resonates with users who support a service aligning with their values. Becoming profitable allows us to reinvest in the business, further enhancing our offerings and ensuring that we can continue to provide a top-notch search experience.
As long as they're profitable I don't mind at all if they stay small. They're extremely useful to me as it is, and their small size means they aren't targeted for SEO nonsense, so their methods to cut through all that still actually work in my experience.
Not every business needs to become a unicorn. Some businesses are better at small scales serving a specific niche, and by their report Kagi seems to have found their niche.
Honestly, I find this whole startup mentality, where you only build a company so that you might later sell it off to some megacorp, very strange and off-putting. It essentially means you didn't care about your product and your users in the first place.
From their site:
Kagi was bootstrapped from 2018 to 2023 with ~$3M initial funding from the founder. In 2023, Kagi raised $670K from Kagi users in its first external fundraise, followed by $1.88M raised in 2024, again from our users, bringing the number of users-investors to 93.
Kagi launched in June 2022 and we maintain a public page tracking real-time Kagi growth and usage statistics at kagi.com/stats.
In early 2024, Kagi became a Public Benefit Corporation (PBC).
If you can bootstrap it yourself then there's no need to do this, but those that bring in investors will need an exit.
Most startups just go through the cycle of cheap and great - hit the profitability button and turn into a flaming pile of crap.
> Every program attempts to expand until it can read mail. Those programs which cannot so expand are replaced by ones which can.
Or the newer version:
> My point was not about copycats, it was about platformization. Apps that you "live in" all day have pressure to become everything and do everything. An app for editing text becomes an IDE, then an OS. An app for displaying hypertext documents becomes a mail reader, then an OS.
So in turn, every product becomes bloatware that needs more money to maintain and more users to get more money.
Businesses rarely remain stable, no matter if they're startups or not. Because that wouldn't make any sense. Either they shrink or they grow. You can call this the law of midrange businesses.
Consider a midrange hotel:
Either the owner cares about his business and continually improves the facilities and the experience for the guests. Soon the hotel will have a good reputation and will constantly be full. So the natural next step is for him to increase prices, because there is the demand and also he has higher operating costs. Repeat this process over the years and the midrange hotel is a high-end hotel.
Or the owner does not care about his business and continually lets things decay and become a worse experience for the guests. Maybe because he wants to save on operating and investment costs. Soon the hotel will have a bad reputation and the owner will decrease prices to attract guests, then further cut costs because cheaper guests don't demand much. Repeat this process over the years and the midrange hotel is now a low-end hotel.
And this happens in all businesses, because in the end they are run by people. If you'd love for companies to get profitable and remain small-ish, then you have to make such a company.
American companies, startups in particular, have terrible support. It's really nice to have actual contact with someone when an issue arises.
Kagi probably have a user base of users who are highly attached to the product’s quality. Kagi could lose most of their paying customers should they ever fall into the wrong hands.
But I’m glad it’s like this. A good old company that just sell good products to their happy customers.
[1]: https://kagi.peopleforce.io/careers/v/125936-core-back-end-t...
Hopefully, now that they've proven that you can create a profitable company from doing this, more will follow and we'll get a proper ecosystem.
Like others, I'd prefer to see Kagi succeed at being the best possible search engine and nothing else, than try and do the Google thing and do everything.
If nothing else, Google taught me that just because something is great today doesn't mean it will necessarily be great tomorrow. I can't get attached.
Also am worried about the move to mail, I already have fastmail, and kagi would need to create a heck of a mail client for me to even consider switching. I'd much rather have a company that does search very well, a company that does mail very well, and a good communication between the two.
And I also have less tangible worries about Vlad's demeanor when I see some of his writing in the feedback forum or discord. It comes across as ambitious but not very circumspect, but maybe that's what's necessary to make it in this sector. I won't pretend to have enough experience to offer much opinion on the matter, all I can say is that its unsettling at times.
However, is there really no other model than these two, i.e. being 100% the product vs. paying the "full price" of the search?
Do people really use web search that little? According to the stats Kagi shows me, I make about 50-100 web searches every day on average.
I guess the problem probably has to do with the value of the GPU cycles being lower than the served content. This is most apparent in the case of AI; e.g. if the mining lasts as long as the session, and the server runs 2 GPUs while the user is only running 1, then you can’t complete the “payment” unless the mining continues beyond the length of the session.
That problem has been solved for anybody who wants it solved, through Apple Pay and Google Pay, or even the built-in feature in the browser for remembering credit cards.
Normal people absolutely would not accept websites hi-jacking their computer to mine crypto and hardware manufacturers like Apple would swiftly implement measures to protect users from those freaks.
And no, there is only those 2 solutions.
Features like boosting the niche forums I browse for search results is just a bonus on top.
I agree that I could go back to DDG and not feel like I am missing too much, but that doesn't bother me.
Plus, it seems like having to login is inevitably makes it easier for them to associate activity with a specific real-like person. It's probably easier for them to associate my activity with me when I log in than it is for DDG or Google to track my activity when I'm not logged in and can't be easily distinguished from the dozen other people that used that computer that day.
Yes, this is annoying. They make it as easy as they can (QR code login, session links) but it's still a speedbump.
> having to login is inevitably makes it easier for them to associate activity with a specific real-like person.
That's a legitimate concern. Kagi added Privacy Pass support to mitigate it
I agree there is not a lot of differentiation between stock Kagi search and DDG. DDG still has a few ads but it's not that annoying, perfectly usable.
Kagi's assistants are pretty interesting though. Recently I asked it to find back a post that I vaguely remembered. It managed to generate a bunch of Kagi searches with different keywords and narrow it down to an old tweet.
It does indeed, as was revealed during the Google Search antitrust case in the DDG testimony.
Wikipedia first. No Pinterest, no W3Schools, no Fox etc etc.
They have a nice list for suggestions.
I stopped using duckduckgo after they censored tankman: https://news.ycombinator.com/item?id=27394925
I found that have two main use cases:
1) Search - It is nice that Kagi search is less SEO spammed. That a big appeal. However, I don't reach for a search engine when I have a topical question anymore. I reach for something like o3 with search, so I don't need to dig through a bunch of articles. Over the last few years my pure search volume has dramatically reduced.
2) Maps - I travel a lot. When I toss a business into the search bar, 90% of the time I want to see it on a map. Google's Maps experience is just far superior to Kagi's. I find that I end up typing maps.google.com more often than I am happy Kagi ran the search.
I don't mind paying for a product; it just didn't work for me.
Side note: I personally dislike that their favicon looks like a "g" for google. It's always confused me.
For my personal workflow, I don’t see the appeal of having a separate setup if I’m just going to use the same underlying models anyway.
2) Yeah this is the one thing I still use google for. Though they did recently update Kagi maps (a weeka ago I think?).
I also find significantly less "vibe written" smut on Kagi. I'm not sure if they are targeting it specifically, or if it's getting corralled by some other metric (like ads+trackers).
Per user it’s about 15 queries per day. I’m sure there will be some that are incredibly active and some that aren’t active at all, but 15 per day seems quite reasonable.
I genuinely thought it would be higher, but I suppose bang patterns don't count.
(Posting this mostly so that people who are curious about subscribing to Kagi can get a sense of how many queries they're likely to need to use.)
Date (UTC) AI Tokens AI Cost (USD) Searches
Jun 2025 0 0.000 141
May 2025 0 0.000 743
Apr 2025 0 0.000 723
Mar 2025 0 0.000 621
Feb 2025 0 0.000 556
Jan 2025 10,692 0.000 1,189
Dec 2024 0 0.000 805It seems that Kagi isn't expecting that things will change much [2].
[1] https://www.wired.com/story/bing-microsoft-api-support-endin...
[2] https://kagifeedback.org/d/7107-microsoft-bing-retiring-sear...
Kagi’s official position is not to support regional pricing (visit https://kagifeedback.org/ and search for “regional pricing” to go through the “Implement regional pricing” thread). The service is probably out of reach even for many people in the first world. Even its family tier is expensive.
Hopefully, when it reaches a much higher number of users, it’s able to reduce prices. Or it can just remain a niche service and potentially be disrupted by a competitor.
5usd in america aren't 5 usd in the rest of the world indeed
I know I'm speaking form a position of privilege and this will be unpopular; but I'm not fond of subsidizing other users by paying premium prices for my subscriptions.
Skipping past the top two sponsored results in DDG really isn't that big a deal and still diversifies the search market from Google's monopoly, so paying the equivalent of a streaming subscription service... idk. None of my friends seem interested in paying for what is currently free (on top of the hassle of being logged in everywhere all the time) so sharing a duo/family account isn't an option for me either. Maybe that's something that would put it in reach for you?
Edit: apparently they've shared what it costs them to provide the service:
> a single search costs us 1.5 cents to deliver
https://kagifeedback.org/d/687-implement-regional-pricing/23
They charge 54€$/year for 3600 searches, which is.... precisely 1.5 cents per search. That sounds a little bit too convenient that the claimed cost price is precisely what they charge
This is either not the cost price or relies on people not using their subscription, otherwise you could never recoup R&D costs. Maybe this is total expenses divided by number of searches being run? I.e., it includes wages for the devs working on new features, their administration, etc. (fixed costs) and isn't actually the expense that each additional search incurs to deliver
DDG displays up to 2 ads above search results (most of the time I get none: trying just now, I saw 1 ad when doing two product searches and one knowledge search). The most favorable figure for 1000 ad impressions is 6$ according to <https://spideraf.com/learning-hub/what-is-the-average-cost-p...>. That's 0.6 cents per ad if the advertising network costs nothing. Advertising on DDG goes via Microsoft, so they'll take a cut: <https://duckduckgo.com/duckduckgo-help-pages/company/adverti...>. Guessing that they take 10% (I expect it's more), you're looking at 0.2 cents average revenue for each search I did just now. (Not a reliable figure but as a ballpark estimate.) DDG must have much lower costs somehow (maybe they just take results from Bing verbatim and have few costs of their own)
I switched from DDG. It saves me more than $10/month, both in time and in actual dollar costs due to suboptimal search engine results. It’s as simple as that.
The jump in improvement from google to ddg is probably bigger than the jump from ddg to kagi though. Also, I switched before ddg and kagi had AI search. Kagi’s AI is a huge differentiator for me, and I haven’t used the DDG one that much.
And that €54/$54 is the price when you pay per year, if you pay per month you’re paying more per search (although at least part of that extra money will go towards handling the payments)
I am tired of being advertising meat, and I'm willing to pay to use the services that do not waste my time.
I'm tired about seeing all those posts about Kagi. I have yet to see a single example where it outperforms Google. People just don't know how to search.
Every time someone claims Google is becoming bad, ask them to share what they're looking for and how they search for it, and where the problem lies becomes glaringly obvious.
My original point is that Kagi is NOT better than Google at search.
Also simply bad faith.
Check my response to this article and the thread of comments that ensued for a pretty recent example: https://news.ycombinator.com/item?id=43829655
I have never ever had a problem with search. If I can't find something on Google it means it simply doesn't exist.
I'm still mad that they took away the '+' operator for that turd of a project. [0]
[0] To be clear, it totally could have been a great project. Early on, there were signs that it was going to be -at worst- decent. But, well, Vivek Gundotra wanted the project to be a big turd, so it ended up being a big turd.
Somehow, Google decided to show me the ferry timetable: https://imgur.com/a/bgFax59
Like, whut?
On AI - I think they have no choice tbh, they need to bring something to the table there. I'm pretty pleased with their AI implementation. In search it only activates if you append a question mark. Their assistant is a pretty good alternative chat interface, it lets you choose the model. I cancelled my chatgpt sub because I can use kagi for many models. It has probably fallen behind the tooling others have at this point though.
> You like the service but don't agree with the company spending money on... t-shirts [...] It smells a bit like you want to control what they spend on?
i like the service, but spending money on t-shirts seems unreasonable to me. I'm sure there are a dozen things which would benefit from the money and time which were spent on manufacturing clothes. From their own blog post [1]:
> The process from here involves setting up a business entity in Germany, so we can import the t-shirts, store them in a warehouse, connect inventory logistics and ship them all over the world. This includes building a website and connecting it to a back-end database
which sounds to me like not the best way to spend the company's time
also:
> why did we go through all this trouble and allocate nearly a third of our investor-raised funds to produce and freely distribute 20,000 t-shirts
all their answers to that question they asked themselves in the same blog post seem silly to me. Go give bonuses to your employees, upgrade devices, make a company event, etc
> On AI - I think they have no choice tbh, they need to bring something to the table there
absolutely not! If i want an AI search result, i go to an AI provider of choice. Again, i don't think that should be their focus
> I cancelled my chatgpt sub because I can use kagi for many models
replace chatgpt with claude and reverse the statement and it will be true for me — i cancelled Kagi subscription because i use claude as a search engine
But from my first days experience I can safely say that 300 searches a month is a low number for entry pricing. And given that I'm in a developing country, the entry pricing is also not cheap enough.
But I'll keep pushing for a couple of months.
2.) Switching to "All Data" also doesn't set the y-axis to zero, but to 6,840.
Having an y-axis set not to zero is in the majority a sign of people who want to inflate growth.
I really like their browser, Orion though. It's still rough, and crashes at times but it feels like a great non-chromium option for MacOS.
They go so far as to have an API to generate cryptographic proof of subscription tokens without revealing your identity for searching when using Tor, etc ( https://help.kagi.com/kagi/privacy/tor.html ).
Beyond their data collection stance, which I believe, their results are better and worth paying for because they don't have all the extra ads and crap shoved in them and allow you to modify your own website rankings, etc.
As long as I cannot access their data, cannot see their systems on a continuous basis, they can collect my data. The same applies to DuckDuckGo and the same applies to VPN services.
So at Kagi they are aware of Tor, they are aware of the Dark Web. If you even talk about this on your website, of course you find ways to circumvent privacy there.
I tested Kagi and my experience is the results are even with or worse than what other services present like Google or Bing or even Perplexity. And if you don't like advertising, just use an ad blocker. I'm so surprised that so many people don't seem to do this and they always complain about ads. Just use an ad blocker already.
So they allow me to modify my own website ranking. Well, if almost nobody uses this website, that's pointless. Besides, if I could modify my website ranking, so can everybody else. And now we are back to search engine optimization.
To be fair I think Google has lowered their position a lot in the results because it used to be that it would be 80% Pinterest results and now even in images it's only a few here and there.
I think they understood that everyone hates that website.
Migrate fully, after Google first search turned into a cesspool of scam AD's and SEO crap.
Local deep searches with ollama are comparable in quality and cost nothing.
Also, half of their premium model list is laughable - 32/70b models can be run locally without an issue.
People truly don't gaf about anything as long as they're getting something for free. Sad.
I assume this is paying users, if this is total MAU ... damn.
PS. I do not mean Kagi is a bad product. I think they're great, actually. I'm just complaining about how poor the reception has been.
And what matters to me is: that it’s enough to keep going, securing the service, and that the number is steadily increasing.
How do you know?
My napkin math says they're in red numbers, but let's see yours! :)
I think Kagi is great. I also understand why most people don’t need Kagi. It’s not hard to see why most people aren’t interested in yet another monthly payment on top of all the other things we’re asked to pay for right now.
Once I tried openrouter, I am not touching any subscription LLM providers.
Google/Microsoft/Apple "all-in-one" personal subscriptions are a different beast, because they charge well below cost.
As much as we like to complain about ads, we also aren't really prepared to fund things that aren't ad driven.
Frankly on the web, for me, ads are fine. For TV less so. I pay for streaming (no ads) and for sport (no ad interruptions) etc.
For search, half the time I'm clicking on the ad anyway.