So rather than getting them to cancel, pause their subscription. You don't have to deal with cancellations, and if/when the user does return, you are one step further than you would be with a new subscription.
Furthermore this generates goodwill, and I'm guessing goodwill has some % that converts to conversions and lower churn.
I know you're being a little facetious but it is actually a benefit. Many companies have implemented subscription pausing to reduce churn. The reason is pretty straightforward good business: it's easier to reactivate customers who lay dormant for 1 or 2 months than it is to let them churn and have to re-sell the product to them from scratch.
Sometimes the simplest explanation is also the correct one.
(historically not so important to companies in practice, but it sure ought to be)
Indeed, this would make me way less annoyed at the thousand and one streaming services popping up like mushrooms after a rainy day.
But the tradeoff would be worth it for sparsely used applications.
For VS code compatible plugins checkout Cline/Roo or just search for AI in the extensions.
Compare to ChatGPT, which is much more expensive, but the value relative to Google is pretty obvious.
On the other hand, the fact that we're having this discussion does point to how difficult it is for Kagi to explain its value proposition and differentiate itself from Google.
As for chatgpt - I'd say its functionality relative to google search is obvious, but not it's value.
It does seem likely though that it's not going to be better for absolutely everyone, other than in terms of having their business model being "give good search results" rather than "give people adverts we can charge advertisers for".
Date (UTC) AI Tokens Searches
Feb 5, 2025 0 64
Feb 4, 2025 0 43
Feb 3, 2025 0 19
Feb 2, 2025 0 24
Feb 1, 2025 0 19
What is the value of ChatGPT relative to Google? It's not obvious to me.
> No ads. No tracking. No compromise. Just deep, powerful search.
So you are not paying for better search but for no tracking and no ads. If you don't care about those, you're not kagi target audience.
Keep up the good work guys!
What I also love is Vlad / the Kagi team's fierce neutrality. For example, there have been complaints about including results from certain indexes like Brave and Yandex, or about suicide, or other political / sensitive stuff and Vlad's response is virtually always a shade of "no matter what, we will display the results because we are a search engine foremost".
Oh and they have built-in CSS injection (under Settings > Appearance) which allows you to hide Reddit's crappy pre-translated search results. You could do that via Violentmonkey / Tampermonkey, but that won't apply to devices that don't have it.
You can also rewrite URL results. So AMP to non-AMP and reddit.com to old.reddit.com (Advanced > Redirects).
Meanwhile Google obfuscates even their divs to make blocking certain results (read: ads) more difficult.
Here's the CSS snippet hiding translations:
/*
Hide pre-translated webpages.
"sri-group" is main result, "__srgi" are sub results.
You can append `:not(:has(a[href*="tl=en"]))` to allow English translations.
*/
:is(div.__srgi, div.sri-group._ext_r):has(a[href*="tl="]) {
display: none !important;
}*I get a lot out of their regex redirect for their search results, notably redirecting reddit to old.reddit -- a lifesaver when searching on mobile.
I abhor sites that translate into English based on my IP. In one case (a job site), I blocked the endpoint for their translation service and that was that.
On iOS there is Modificator which allows to inject CSS and JS:
- Ranking results from specific websites has been well referenced in comments here. I love always knowing if something is on archive.org and wikipedia by having those results come to the top. I also rank certain sources of medical information up and down based on reputability, basically overriding their SEO nonsense.
- There are subtle indications for sites that have a high number of ads and trackers, allowing me to opt not to even click on those results.
- AI summaries and answers are not on by default, and simply adding a question mark to the end of my search allows me to get an AI generated answer to my inquiry. I've found these to be very good, but I don't always want them so the control is great.
- Marketing and ecommerce sites seem to be aggressively minimized, which makes the internet feel less like walking through a mall. I only really go to Google if I am shopping for something and want those kinds of results, but this is rare.
All of this makes for a much better experience of the internet overall for me. The reduced cognitive noise is well worth the $10 in my case.
I can't speak to how it preformed in non-English content, so you may be well served by using Google for German content in that case.
What makes Kagi great is that they let you customize results. I've pinned wikipedia, for instance. Google first throws AI slop in your face (with no way of disabling it), followed promptly by (presumably also AI-generated) blog spam, Pinterest links, and other useless garbage that I can't filter.
fwiw, I search in German every once in a while and the results are a lot better than Google (in the US, anyways), since I don't need a VPN to get "good" results and have a quick toggle button for my location built into Kagi.
Also, as a company, they seem great: They are neutral, run as a PBC, are very open and transparent about what they offer and why it costs money ("no BS", if you will), are receptive to feedback and do consumer-friendly stuff like this change.
I was sold when it helped me uncover pages I'd never read before about an extremely niche local history topic.
Really, it’s even better than that given the full feature set.
What I love most about this fair pricing is it makes it was more appealing to encourage my friends to give it a try. Thank you, Kagi!
I would really struggle going back to not having bad sites suppressed in search results!
Good job Kagi et al!
Hopes that netflix or any other provider will implement this are small though. Because it's free money when someone pays for service and does not use it.
https://entertainment.ie/on-demand/on-demand-news/netflix-ac...
About the fair pricing: Would love to have this also for my car lease ;) But more on a weekly bases.
The flip side of that is that only a small fraction of their members could actively use their memberships or they wouldn't have enough space. The active members get their membership effectively subsidized by people who don't use their memberships.
Apparently up to 50% of a gym's sign-ups happen in the month of January due to new years resolutions, and January/February are the busiest months as a result, though the majority keep their membership even after their resolve to go tapers off.
Right. This is the sort of pro-consumer practice that is obviously morally right, but will not be widely adopted without consumer protection laws. Outside of small, niche businesses like Kagi, there is no pressure to treat customers with respect.
I found out about this because I noticed our Slack bill was quite a lot lower over some Christmas/January period. It was because so many folks were away, and so they didn't charge us for seats that were inactive for > 30 days.
Kagi in the other hand is useful.
Probably that's why Netflix has to play hardball with their customers, chasing their money hard and strong, pushing them around, not Kagi? : )
A research paper from a few years ago introduced the concept of “customer inertia.” It found that users tend to overestimate their difficulty in unsubscribing from a service. In other words, when a subscription includes auto-renewal (or a similar feature), a significant portion of potential users will choose not to subscribe because they fear they won’t be able to cancel if they stop using the service.
According to the study, this affected about 30% of users. So, could offering something like fair pricing reduce this barrier and increase new subscriptions by 30%? https://bfi.uchicago.edu/insight/finding/sophisticated-consu...
What does work for me is when the service's docs have a very clear page on how to cancel the service without having to talk to someone.
The disconnect between the researchers and people's actual estimations is that "cancelling a service" is much harder than the couple button clicks it usually takes. You have a structural problem: If you don't use a service, you don't spend a lot of time thinking about it. It's easy to cancel something if you make a conscious decision to stop using something. But if it just gradually falls out of use, your only reminder that you should cancel it are your bank statements or the occasional payment reminder email (that some services avoid sending for exactly this reason).
I would love to see the FTC mandate a policy that prohibits automatic renewal billing if the service hasn’t been used for some time.
Obviously some services like insurance or storage don't work like this, though. I don't want to use them, but I want them to be there if I do need them.
- Act on customer complaints (or consumer protection organisation complaints)
- Proactively investigate and check
- Require businesses to submit proof that they follow the regulations e.g. test results
I’m sure there’s other ways and you can do one or more of these things to ensure compliance. It’s really context dependent on which methods one would use.
This was an exact point I raised when they attempted to charge an expired card twice and then sent my bill to collections. The gym staff admitted to remembering that I attempted to cancel because I was moving to a place with no Anytime Fitness locations; they refused to let me cancel my contract early without me showing them my new lease, which I didn't have yet and wouldn't have until after I had already left my old city. They also surely had electronic records confirming that I had not set foot in an Anytime Fitness since that time - or else, no ability to prove that I had set foot in one since that time.
That they had the nerve to not only keep charging my card but send the progeny of their multiple degrees of utter failure to collections is exactly why they never got a dime out of me. If anything they owed me money, not the other way around. That hundred or so dollars has since rolled off my credit report, but until then I wore that delinquency as a badge of honor. That shithole of a company can shove it.
...anyway, that'd be the way to enforce it: by checking access logs to see if the customer actually used the service. Don't have access logs? Well then, you know the saying: customer's always right.
We also have legislation that provides warranty on electronic devices and household appliances, everything really, except things like cars and boats etc etc, for the reasonable lifetime of the product. So a cheap washing machine, three to five years would be reasonable, an expensive unit? I want that to last six to eight years. An expensive fridge, at least ten.
I also thought for a while that things like ChatGPT internet search or perplexity would replace DDG and Kagi, but, so far, I just want slop free sources to back up the slop I generated purposely in R1.
Their Quick Answer feature does an AI summary of your query results. By default it shows up automatically when it has high confidence. You can disable this in settings or force it to show up by adding "?" to your query.
https://help.kagi.com/kagi/ai/quick-answer.html
If you want to jump into an AI chat session, you can add bangs to your queries. "!expert" launches a top of the line research agent and "!code" is good and software development. Both of these use the underlying search engine to get current facts.
Kagi even maintains their own LLM benchmark to monitor how well different models perform. They occasionally swap out default models to keep performance SOTA. You can specify a specific model if you want.
I also really like this model for subscription services in general. It would be nice to, say, not be billed by Netflix (though really, I’m looking at Paramount+ or Peacock) for months when you don’t use the service. It’s the kind of thing that wouldn’t be hard for companies to implement, and could potentially be regulated into existence everywhere by bodies like the EU or the CA state government.
Now, wouldnt it be even better to implement usage based pricing with a maximum that's equal to the current subscription?
I dont know the Kagi details. Say you pay 10/month and each search is $0.02. You pay max(search_count * 0.02, 10).
I guess the logic is much simpler for their current system. It's 10/month, period. Then if you didnt search anything you get a refund. Instead of tracking and calculating usage.
However, usage pricing should be more enticing for casual users. With the statement refund for 0 use, there is now an incentive for infrequent users to NOT to use the product.
However, Kagi search unlimited is $5/mo. And, especially because of the whole "payment providers taking their cut" thing, I'd argue for services that are $5/mo or under, metered pricing doesn't really make sense.
Finally, I'd argue that this approach has other advantages over metering, even for higher-cost services:
* It's easier for devs to implement. Just one search needs to be recorded a month for an audit trail, rather than all search history.
* Keeping a search history for users is not needed at all, really (or, again, at most, one search per month). It's much better for user privacy.
* Most countries/states would have much more luck passing legislation forcing companies to implement this than metering, as well as enforcing it.
The academic lens is like Google Scholar, but better. The papers it surfaces are simply higher quality.
Otherwise, append your query with a question mark. The baby AI will do what Google's tries to do, except with a little more skill and better citations.
Most broadly, however, search. It's kind of wild but I forgot that searching the internet used to be fun. Kagi made it fun again.
Also I guess part of this is probably the option I used to give higher priority to some websites like python org.
When I subscribed with Kagi, I was so totally pissed off and stressed by using Google where you will now have crap and unrelated ad links everywhere on the page. And in addition often first link that are garbage Copycat of principal websites. For example, for python, when looking for a module documentation, the official doc is the best but there would be hundreds of ad filled shitty pages that would appear first.
I haven't used Google in many years so I can't directly compare the search quality, but Kagi is good enough that I've never had any reason to try something else since I've started using it.
Edit: I also use the !w Wikipedia bang constantly, I forgot that was a Kagi feature and not my browser. Obviously Kagi is not the only search engine with this feature though.
Also, it's built in that sites with a ton of ads are down-ranked.
Currently they charge 2.5c for an API search. This is between 1,000 to 1,000,000 times more than other companies in the space charge.
AI systems need to do dozens of searches for every question to get good results and kagi's results are really good. But not 1,000,000 times better than the competition.
At that point all the special sauce Google et al have spend decades mastering will be worth as much as expertise in analogue computers is today.
So it's either unlimited or nothing. But since I know Google's search operators well I don't have trouble finding things if they exist so $10 per month is hard to justify. Plus, you're anonymous with Google but you're not anonymous with Kagi since you pay them.
But Kagi can be good for tech illiterate relative you want to shield from sus sites.
Glory and liberty for Ukraine!
Thank you for your support.
- Zac
* I don't trust the product's claims. Sure, privacy and user-centered results sound cool, but literally every company on the internet claims to cater to the user and value their privacy. Kagi can apparently afford to be more specific than usual, but how binding is that? I don't know, I'm not a lawyer and definitely not versed in US/California law, and given all the obviously exaggerated claims in this domain by all kind of actors, I can't give it much credit. I guess Kagi has to pay for the whole industry's decades of malpractices in this regard and that sucks, but I guess you could do better if you opened more about your
* I don't trust the product's ability to stay around. Startups come and go, and I'm not subscribing to a paid service and switching workflow without a reasonably solid belief that I won't have to do it again in a near future. Your new pricing policy actually helps quit a bit in this regard, the other bit requires you to actually stand the test of time, so just keep on doing your best I guess.
* Pricing has is shown excluding taxes. I'm not going to figure out the US tax system just to know how much I actually to shell out, and I'm not paying if I don't know how much. In Europe, VAT is around 20%, so it's a pretty significant figure, that would be 60 bucks a year for the Ultimate plan. I don't have the slightest idea if that's the order of magnitude expected in California. Have your lawyer or accountant figure it out, because I sure as hell am not. Allowing me to pay in euros would also be a quite large hurdle removed, for similar reasons: exchange rates fluctuate, banking operation costs fluctuate, and even if I can work it out more easily than US taxes, I'm not going to do because this should be your job, and whatever figure I work out will be obsolete by the next time I'm billed.
I also generally have this mindset, but I've come to think of it through the lens of me getting a better experience for a while and going back down to what I had before vs having never had that better experience.
Before I paid for Kagi I worried "what if it's great and then they go under?" But then I'd just go back to Google and move on, having never had that better search experience. I guess it's kind of like "better to have loved and lost than never to have loved at all?" Except that quote seems a bit over dramatic for a search engine...
Not user-centered (Google) : they don't give a dam about user feedback, the results provided are based on how much money they can bring to the company through ads/affiliations
Why does it matter for a search engine? It's not a tool in which you store your data and need to be stable.
You use it now, when it's available. If one day it stops being available, or stops being good, then you stop using it. Nothing lost compared to not having used it in the first place.
I have been using Kagi since mad way through my wife's pregnancy, and my son is now not far from a year old. Notoriously, the moment the internet gets any sniff about impending or recent parenthood, every advert becomes about nappies etc.. But I haven't had this problem at all. I've done hundreds of searches on everything from toys, nappy brands, to newborn medical stuff, and my adverts stayed firmly child-free.
It wasn't until my son was about 6 months old that I saw any adverts at all, and I'm pretty sure that can be traced back to a FB post (I don't post often).
You pay the amount of VAT based on where you live. I agree it would be better to display that on the pricing page though.
As for the privacy claims they have a fairly easy to read privacy policy that goes into details about what they do and particularly don't do with your data. There is no vague wording to hide behind.
Real "fair" pricing just charges per request, and has the per-request pricing progressively go down as they reach various thresholds. Preferably with a free tier.
To ape someone else’s lament: I can’t take advantage of this because I use it daily.
Whenever I mention Kagi is actually better, someone will claim the opposite.
So yesterday someone here said something along the lines of: "With apologies to Bill Buxton every user interface is best at something and worst at something else".
So I started looking on Kagi and only found a few results, even if I took parts of it, but I narrowed it down to that the original must have been about "every input".
Guessing that Kagi had excluded a few results so I tried in Google (Googles usual problem is adding things I never asked for and I wondered if Kagi had become overzealous or something).
So here are the results from Google:
https://www.google.com/search?q=%22Every+input+is+best+at+so...
For me Google says:
> No results found for "Every input is best at something ".
> Results for Every input is best at something (without quotes):
https://kagi.com/search?q=%22Every+input+is+best+at+somethin...
So now, while Kagi has always been a lot better at not including unwanted results, it now seems it also has a larger effective index than Google.
Kagi attempts to only provide results it thinks will be relevant. While I liked the accurate results, I was frustrated when none of the 5-10 results was what I was after; at that point the UX is to type a new search term rather than simply scrolling further (I prefer the latter).
One other small downside is I slightly missed google's 'WebAnswers' (certain google searches will display images and summary info for the search term, rather than strictly results). WebAnswers were handy on super quick searches for, say, a particular car or aircraft model). I didn't think I'd miss this, but I did, although it was very minor.
I put money into the account, you bill me per search - pre-paid usage based billing is the only way this can ever be "fair".
Now I use the unlimited plan and so I search first, spellcheck later. Or sometimes it corrects it for me.
This only works at the extremes of volume. If you're targetting very-low use users, or enterprise, you can price per search. In between the frictions just don't make sense for any sensible target market.
I suspect this would also work great for streaming services, like HBO or Netflix. Rather than paying for unlimited use on 6 platforms OR spend fortunes on pay-per-view on yet another platform, just reward your most loyal customers but keep the door open for incidental users.
Last time I checked in on this, Kagi was bootstrapped. The single biggest motivation for me to make a bootstrapped business, is to make an ethicals busines.
This includes ethical pricing, ethical communication, and ethical UX.
For around the same price, I can stream millions of songs, or stream thousands of high res videos, or subscribe to both premium e-mail and a premium task manager.
What makes web search so expensive?
If what you mean is you can pick from their library of millions of songs, Kagi sounds like an even better deal. For $10 a month you can search 400 billion web pages.
The first time I encountered this was with Slack—they only charge for active users.
We follow a similar approach with our products : 1) PodcastAPI.com - If no API requests are made within a month, the user pays $0. 2) website Premium Membership- rather than forcing users to pay a monthly fee, we allow them to buy a 2-day pass with one time payment (default option) - listennotes.com/premium
Caveat: customers will demand more. soon, they’ll request for hourly fair pricing - don’t charge me for those hours that i don’t use your service!
Also, there was significant latency in searching compared to Google.
Out of curiosity I just tried the same search, and all the results I get are for .ie domains, so it looks to be working correctly.
It’d be great if they extended it to refund $5 for anyone on a Pro or Ultimate plan doing less than 300 searches in a month, too. (I pay for ultimate and would still be very happy with that gesture.)
Kagi also actually has a business model. Mozilla has a teat that a US Court might order removed from their mouths soon as a possible remedy to sanction their Mommy in an antitrust suit; and looking at their 990, things are not looking particularly good for them if that happens.
I think it’s a good balance between locking the user into your product and dealing with the cost of a constantly evolving service.
Bonus is you can query multiple models at once, including local llama.cpp/Ollama models. I use it with the Claude and OpenAI APIs, as well as local Mistral, Qwen, and DeepSeek models.
[0] https://docs.openwebui.com/ (one liner if you have `uv` installed)
I’m not a Kagi subscriber though. The USD 150 and USD 216 a year prices for family duo and family are quite high for many geographies. Hopefully Kagi scales its customer base and is able to provide affordable plans.
A warning from someone who forgot to disable their subscription for 18 months before realizing what they lost.
More annoying to me is that you have to use up your credits before cancelling your sub. If you have credits and you cancel your sub, you lose the credits.
The AI answers all have references too.
It is paid service, what is the other option? People that don't use the service being happy with it?
> because their brain wants to justify them paying for search
It is search engine, not candy crush. No one wants to pay for searching, if they do it is because they find it useful. It is not their brain gets a shot of dopamine every time they do a search on it.
I don't know why you're throwing this out here without anything to support it.
Admittedly it's a tough current to pitch yourself against, that search should be a paid service. But that's mainly because the best advertising company in the world is leading the charge on the other side.
The truth on who's more delusional appears murky to me...
This is _really_ weird marketing, in that it implies that previously the pricing was _unfair_. That's not an idea you generally want to put in your customers' heads.
Imagine the balance of revenue from non-returning users (think fitness) vs very heavy users, and finding a way to keep both parties happy. And the implications it has (those "paused" users still count towards "onboarded" users).
Major props Kagi team, or who-ever pushed this idea!
That reminds me, I need to cancel my 24 Hour Fitness subscription.
edit: use of Yandex on the other hand.. yeah that's a no go for paid subscription.
You don't pay some heavy license fees for a local installation anymore, but get a login where you get billed a specific amount if you use it that day/week/month etc. I was pretty sad when I saw how it got implemented...
Maybe a threshold of a dozen searches or so before the subscription fee kicks in.
I use kagi hundreds of times a day, so it's not something likely to happen soon, but I'm still curious
There are more tools to understand data and squeeze every penny from customers.
It's admirable when a company isn't trying to bleed customers dry.
I liked it, the results were good, no ads, gave me access to Google without being tracked.
I would pay for that, except they block Tor, and I normally use Tor.
We do not block Tor - in fact, we recently launched our own self-hosted Tor node[1].
We have had problems with GCP blocking VPN and Tor traffic (mostly the former) when we have made zero configuration to do so. It's quite frustrating, and we have been working with their support to improve this generally.
Haven't heard anyone having issues with Tor since we set up the node though :) If you give it a try, let us know how it works for you.
[1] https://help.kagi.com/kagi/getting-started/accessing-via-tor...
Hint: you can use Google in private mode. And unless you block all trackers almost all sites will still use analytics so Google knows what you read.
I think we are entering an era, where these things will just not be available to people in developing countries, as they cost more than they can pay. Especially taking Ai into consideration.
However, we could start paying fair prices for produce like clothes.
Search = advertisements
Kagi team, folks say you have a great product, but if you don't pay attention to small issues like this one, you are bound to lose some of your goodwill.
> also they blocked my vpn's another big nono
Argh, I am sorry to hear that. We do not block _any_ VPN, but have been struggling with our hosting provider (GCP) to not do this. We are working with GCP support on this - if you give us a chance, and encounter any issues, please reach out to support@kagi.com with details and we will see what we can do.
Also, see my other comment in this thread about Tor, if that is helpful for you.
Imagine Tom Cruise in a variation of Minority Report. As he enters the shopping mall, the onslaught of cognitive infiltration envelopes him. He's not there for recreation, nor to evade or investigate anything. He knows why he's there. Or, he did know, but now finds himself trying to remember as he fends off sleazy desires for strange things. He knows he doesn't need more ugg boots, the unworn pile in his closet and fact that he's never worn boots of any kind a testament to this. He knows a new car won't reignite the wonder of his youth or make the foggy shores of a moribund sea glisten with golden light. He couldn't afford it anyway. Despite a lobotomizing decade of overtime and side hustles, the red queen always stays ahead. It's those damn conversations with the pariah professor.
If I didn't waste my time with her, my social credit score would expand and I could afford the newest virtual vacation to the green place they say existed before Amazon bought the planet. That hag is oppressing me, damn her!
Tom was different though. Somewhere in the vestiges of his mind he knew this was bullshit. She was no hag; she was beautiful and fascinating and wise. It was her and only her that made him think again, to contemplate meaning, to ask forbidden questions, to feel.
"It's just the mall, stupid" he remembered. The enormous image of an inflamed scrotum foisted itself onto his entire being, gracefully rotating to show all angles. That's right... He was just slipping into the drugstore with the sole purpose of buying antifungal cream for the persistent case of ringworm he contracted from that robotic concubine store.
He was becoming disoriented and dizzy. Boundaries were beginning to dissolve and he knew it was time.
An androgynous figure in full lotus hovered before him, emitting a calming hum. In its halo could be faintly seen a scrolling index of the stock market. "Do you want sanctity of mind? Is it time for inner reflection? Do you need focus?". "Buy Now Pay Later!" it hissed.
In a whirling, scintillating carousel of nausea and mumbling faces he lifted his wrist, touching it to the NFC receiver on the hovering being's pulsating third eye.
And suddenly, as if waking from a nightmare, he was human again, with will and self definition.
Compelled to move quickly, he knew there was only 20 minutes, and his balls were screaming.
What a bold and genius move.
No thank you.
I mean this is great. But how are they resisting the global trend to be an advertising influenced portal? How are they not adapting?
1. We have a diverse set of upstream sources. It would require all of them to be "compromised" for our results to completely tank.
2. We have a "crawler-lite" that solely collects info on page quality - number of ads/trackers detected, page speed, etc. - and we take that into account when ranking the results, generally nudging them down the page if "cleaner" ones can be found.
3. Our sole source of income is our users, and we actively respond to impacts in search quality that our users submit to us on our feedback form. If they are not happy with the results quality, then we are out of business.
Put simply: Our technology and business model is completely aligned to resist "enshittification". We have no reason to bias results or "sell out" to anyone.
With the momentum we've gathered, we are also taking first steps to building our own full scale index[1], both as a valuable contribution to the ecosystem of search and a contingency to reduce our reliance on 3rd parties.
[1]: Help us build it! https://help.kagi.com/kagi/company/hiring-kagi.html
I'd happily pay a one-time fee for a 1000-search package that would be added to my 100 free searches.