Just this morning I bought a month of this to try it out and honestly, I’m already sold on it. The concept is a lot like Kagi: a paid search engine without ads and some great customisation, such as prioritising, pinning and blocking certain sites. Main differences: Uruky is EU-based, with even its results only based on EU-based search indexes; its creators do not want it to turn into a sprawling service provider like Google or Kagi; and it’s AI-free. Here’s an interview with Bruno, half of the husband-and-wife team behind the project.
Though this would be of interest to some of the users here.
Interesting to see new contenders in this space. I noticed that Bruno mentions offhand the absence of LLM features but doesn’t really treat that as a selling point, which is a missed opportunity if you ask me. (See for instance Kagi Criticism on the indie web wiki.)
Yes, I agree that going all-in on being the AI-free option would draw in more people. Just see the number of people in the Kagi forum saying they’d even be prepared to pay more for an AI-free tier - they’re just waiting to be snapped up by a wily AI-negative competitor.
For the moment there are a few drawbacks to Uruky that I’m happy to endure, knowing this project is only a few months old.
images don’t didn’t pop up in a nice little subwindow when you click them. You’re taken straight to a new tab with a cached version of the image, which doesn’t always exist. *added as of 1 August*
there is was no ‘verbatim’ mode like Kagi has, which I had set as the default because I don’t like my results muddled with stuff that kinda sounds like what I typed. *added a while ago, see Bruno’s comment below*
That’s it for now. As for search, I’ve not felt like anything’s obviously worse than what I had before.
edit: Bruno has clarified that AI is not an area of interest for this project and both of the drawbacks above no longer apply so I’m a happy camper. Paid for a year of Uruky!
Wow, I was quite quickly bought into trying it out. I still wonder if it will perform well, and if it’ll live, but after hopping across a couple of the search options like Qwant, SearXNG, Startpage — I’m willing to give a paid search option a chance. I wasn’t with Kagi as I found out about them after they went sour and people started being on the fence about it.
Feels like I agree with a lot that the lead engineer Bruno had to say about privacy and mechanics too — Mullvad’s payment system is a wonderful thing in my eyes as well! I also support the recent push for Europe to become self-sufficient and autonomous in the tech space.
Thanks for sharing, this may just be a diamond in the rough!
Hello, there! I’m Bruno, Uruky’s co-founder. A customer sent me to this forum that I’m delighted to learn about. While I plan to lurk and learn and contribute beyond Uruky, I wanted to take this opportunity to clarify a couple of things, here:
I believe (can’t be certain, thankfully) that @rmf has emailed me before and that’s why we have the “Verbatim” mode as of a few weeks ago.
The image popup/gallery thing is something that’s in the backlog, but we’re still focusing on our indie/small-web index and supporting more independent search providers to give people more control and choice.
We were hoping the “Not another AI tool” header/point on the landing pages was enough of a statement to make it clear we think it’s a selling point that we won’t add LLMs/Gen AI.
That’s it. If you have any questions, find anything confusing, or are running into any trouble, feel free to DM me, email me, etc.
Hi Bruno! I’m a current Kagi subscriber that’s been considering switching over due to Kagi’s use of AI and their use of Yandex with no option to disable that or other indexers. And I have no need of Kagi’s larger ecosystem so switching would be cheaper too.
In this message you said “we think it’s a selling point that we won’t add LLMs/Gen AI”, but just to be clear, that’s not a promise Uruky makes, right? For example, on the page comparing Kagi and Uruky it says “Uruky doesn’t use AI (for now, we’re still unsure on how to fit it sensibly)”.
Thanks @thepaperpilot! I’ve mentioned this in a few interviews and public places before, but I suppose we still have some older “less clear” messaging on the website we need to address. To be clear, “AI” means a lot of things that aren’t related to LLMs, which we could think of being usefully applied (NLP or ML for NSFW filters or better abuse pattern detection, for example) — though we have no plans for that yet as it’s not clear how it could be respectful, etc. It’s why that messaging was initially like that.
That being said, what we are certain we won’t do is provide generative AI/LLM-based features. We’re not interested in that. We’ll brainstorm some “marketing-like” (read simple) ways of conveying that information on all our public pages. If you have suggestions, I’d like to hear them!