I want to start by inviting everyone who are looking into this post to check the git commit linked before reading and after that read their current page, it will make more sense.
I think most people don’t understand the fact that Kagi is meta-search engine which literally collect the results of other search engines and display it and add very small amount of results from their index(tecilis).
They are starting to act shady this year(check the attached git commit to see their changes to their documentation page).
People basically pay them to search Yandex and Brave for them.
Before anyone say but privacy, they are closed source and there is a better open source meta search engines which actually respect your privacy and need your donations more than Kagi shareholders.
Open source alternative:
It’s not fair to call Kagi just meta search engine. As you mentioned they do have their own index, it’s just that they are complementary. They even show you how much of the results are unique results from their index. They’re very open about it.
there is a better open source meta search engines which actually respect your privacy and need your donations more than Kagi shareholders.
Privacy I can definitely agree on but the actual search result is not better. I think it’s safe to say that if you’re paying for Kagi, you know what Searxng is and can spin up one if you want. I’ve done it myself and it was clearly not better.
And if you think about it, if you self host Searxng, that means all the search query + your IP goes to search providers which kinda kills the point, unless you use VPN or Tor, both of which have their drawbacks.
If you use hosted instance, you have to trust them. Being an opensource project doesn’t mean a lot when you can’t go through the source code that is actually running.
At some point you have to trust someone. I happened to trust Kagi to some extent.
I tried SearXNG but the UI/UX was so shitty that I’d honestly rather pay for Kagi. Lots of people love it because it’s open source and are willing to 100% look past all UI/UX problems, which is a very common problem with open source in general.
Also, the actual quality of the results was garbage compared to Kagi.
You can change UI theme on SearchX
I know, and all the options look pretty bad IMO. That also doesn’t fix UX problems.
And I’ve found SearxNG to provide fantastic results, though I’m using it self-hosted so maybe that’s the difference? I started using it as a test about a month ago, and I’ve converted by default search engine on my machines to it because it’s delivered excellent, un-cluttered results for anything I’ve looked for on technical and non-technical searches.
I do and I like the results? Like, I get how we have to be cautious with closed-source private services. This just reads to me as clickbait to put down on people that use kagi lol. Also searxng isn’t really private either as you can’t be sure what your instance is actually running, and self-hosting just links everything to your IP as other people mentioned…
SearXNG’s result quality and blocking functions are much worse than Kagi’s in my experience.
edit: Downvoting personal experiences. Lemmy going strong again. ;-)
Join my instance, and you’ll never see another downvote again.
Why in the world would I not want to see downvotes?
You can actually hide them in Voyager for example. Don’t need a special instance for it. You can also set the app to combined instead of individual tally.
To answer your question, upvotes and downvotes are fleeting on lemmy and can be oft meaningless. You could still sort content in posts by metrics, but hide votes, and get insight out of it.
This instance actually drops downvotes before they hit the database, so unless another instance tracks users on different instances, they simply disappear for people on this instance.
Because downvotes are lazy commentary. I’d rather judge for myself what constitutes a bad take and use words to encourage or debate. I don’t need a bunch of angry keyboard warriors poisoning the discourse with voluntary polling.
Plus, seeing a bunch of negative numbers doesn’t make anyone feel good. I would rather Lemmy be a better place than Reddit.
I like this perspective of yours, and while I don’t think I’ll join you on your instance (I take it this was an open invitation, yes?), I’ll look into ways to hide votes on my client.
Although I do wonder, would you prefer everyone to return to long strings of “+1” and “agreed” posts?
Not in particular, but at least I could ostensibly set up a filter (or automod) that hid or removed low-quality comments like that. Removing downvotes is kinda the same effect.
I’m not saying voting should go away entirely. This instance still has upvotes, after all, but Lemmy will just turn into the cynical, pessimistic, self-fallating shithole that Reddit has become if we don’t do anything differently as a community.
(And yes, it was just an open invitation—a reminder, if you will, that The Fediverse is a cool place where you have choices regarding how you experience it.)
+1
It amazes me to understand that your reply violates your instance’s ToS:
No self-advertisements or spam
(Still, thank you: TIL.)
Kind of ironic, that. I do not apologize. 😆
Shameless too!
Well, yeah. It’s pretty much the same as SearXNG. Except paid, closed source and needs an account to function. The results are better though, I’ll give you that.
…and the pricing is absurd. I’d like to disable all the AI bullshit and get a discount.
If it was self hostable, I’d always recommend it over SearXNG. Let’s say that it’s a better but closed source and paid SearXNG alternative.
needs an account to function
How would they be able to provide a way for you to filter and/or downrank certain (known “less helpful”) domains without you having an account?
You can store them in the browser locally.
So you’d need to set up your whole block list on every single device you use?
SearXNG generates a URL that contains all the preferences you’ve set, so it’s a matter of copying it and setting it as a default search engine on another device.
For a privacy focused search engine. At least the option would be nice. I personally just use my PC and phone, so I’d only need to set it up twice. Additionally, I imagine an export/import function wouldn’t be too hard to set up.
Convenience or privacy? Choose.
Why not both?
Sure.
Self host your own database and synchronize it that way.
Sync it with firefox?
If they have a great meta-search algorithm, users would be able to search without an account and see how great the results are. Then, when a user wants to personalize ranking and block sites, they can create an account.
I always assumed that they make you create an account to track search usage and cut you off once you hit the free tier limit.
People in the tech community who care about their privacy need to understand that if a product is not open source it should NOT be trusted. Transparency is a necessary factor in guaranteeing privacy, otherwise we cannot see and prove if a person or company’s claim to privacy is true.
But be careful, being open source does not inherently make it trustworthy. It could be doing something nefarious that just hasn’t been caught yet.
Indeed.
However, proprietary code cannot be trusted to be trustworthy. FOSS can it’s just not guaranteed.
It’s not impossible to analyze and test compiled binaries.
While not theoretically impossible it’s impractical and unheard of for multiple trusted independent audit companies to be commissioned on an extremely regular basis.
Trustable proprietary software is like nuclear fusion as an energy source. Theoretically possible, but we sure as shit don’t have it.
Now you’re not going to believe this, ladies and gentlemen, there was a time - not too longer ago in fact - where people called these metacrawlers, and they were beloved. People would say “Dogpile” yes, that might help, because - bare with me again - search was shit. But, comparing lot of different shit sources would generally give you ok information, and generally reveal at least a lot of different sources which could help find and gather information.
Now I’m not calling Kagi reputable in anyway, but if metacrawling helped with the situation before, maybe we should try it again.
I don’t quite understand how this is an issue.
I mean personally I’m happy with anonymised ads from DDG in return for anonymised Bing searches (+their own guffins) but I think it’s fair enough, if you really want to see no ads, that some turn to Kagi.
Basically, you’re surprised that people will pay money for someone else to host and run a service for them instead of hosting it themselves?
Kagi has good search results and they are presented well. It also has some useful features like forbidding certain sites and prioritizing others. I like that by paying I’m the customer and not the product. And their “small web” initiative is commendable.
That said, I’ve been a customer for nine months on an annual subscription, and I will not be renewing. The first reason is that I find them just too expensive for what they do. The second is that, even being that expensive, they’re not breaking even. That undermines my trust in their future as a search engine and makes me less interested in paying a little extra for a good cause.
even being that expensive, they’re not breaking even
huh, I guess starting their own T-shirt business wasn’t that great of a move after all
It wasn’t even a business, it was just a huge purchase of tshirts for marketing, apparently, right at the beginning of their startup. Like, I have no idea why you would think distributing t-shirts would build your search engine’s business, let alone doing it when you’re trying to get your technology sorted out and running.
“Hey, neat shirt, what’s Kagi?”
“Why, it’s a search engine you pay for that searches other search engines and aggregates the results for you!”
“For realz, where has this been all my life?!? Sign me up, my brother!”
https://blog.kagi.com/celebrating-20k
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. So, we basically ended up owning a merch production operation end-to-end, just so that we could ensure premium quality of these t-shirts!
like… why
That’s hilarious.
If they’re indexing multiple sources and then making a custom experience for the user - based on multiple sources and user input. Then that to me, sounds like a valid service to to allowed to sell? I struggle to see the issue here?
\not a kagi user.
deleted by creator
I think most people don’t understand the fact that Kagi is meta-search engine which literally collect the results of other search engines and display it and add very small amount of results from their index(tecilis).
People basically pay them to search Yandex and Brave for them.
You’re enticing that they ‘do very little’. But they do search multiple sources at once, providing a better choice of results for their users in addition to allowing users further customizable filters. If they’re doing this then they’re doing more than just the standard ‘Yandex/Brave’ search as you describe.
I do however, agree that they should state that they are a metasearch engine if that is their case. Also just seen their pricing - ludicrous!
Kagi doesn’t hide that they use API calls to multiple sources for each search, they are fairly upfront about honestly. The benefits of use Jagi IME are the results are great, the site is fast and gets out of the way, it’s fairly affordable for what it provides, and the goals of the company is in line with mine (namely to find a thing I’m searching for). They are well funded enough to give me confidence that I’m not going to have to configure yet another search engine, and the integrate into pretty much all my access points easily as a default search engine.
I have seen no reason to think they abuse their position to impact my privacy, and bring closed source does not automatically make them evil. You included no alternatives that are open source, and the ones I explored were either difficult to get setup, required me to run something on my own infrastructure, or didn’t provide the integrations or results I expect. Kagi does.
Kagi isn’t perfect, and there are a ton of suggestions on their feature tracker that users rightly want implemented (including open sourcing more of their code-base). But as a paid search engine that makes me not the product, it does that job well.
they are fairly upfront about honestly.
as long as you don’t really want to know which search engines they use https://kagifeedback.org/d/5185-more-detailed-information-about-search-sources
In the end it doesn’t really matter. What matters is that what I get as a result is relevant. No ads, no 3rd party interests, easy filters, customizable and useful output.
What are they starting to do that is shady and why do you think don’t respect privacy? (I couldn’t see anything on the document you mentioned)
deleted by creator
They deleted the fact that they are a metasearch engine
The diff you linked to has the new text of
Our search results also include anonymized API calls to all major search result providers worldwide
You’re being very disingenuous. Actually, no - your statement above is flat out wrong.
Can you give us some example of better open-source product?
deleted by creator
*NG. SearX itself is dead.
deleted by creator
OK, and some others alternatives maybe? Just to get a clear picture of what is out there
they are a metasearch engine
You do know that they have several own indexes as well?
People basically pay them to search Yandex and Brave for them.
I think this is a fairly benign move with a valid explanation that has been taken out of context in order to be more inflamitory
deleted by creator
They updated their description of how their search engine works. The new description is still factual, while being slightly easier to understand. The new description also has some fancy marketing flair. They still make it pretty clear that they use other search engines via anonymous calls.
I wouldn’t call that shady at all… Unless maybe I’m completely missing the point here.
