I've been using it for a while as my main browser and it suffers from all of the usual issues that early development stage browsers have: bugs, memory leaks. The bugs are nasty as well, I'm sometimes forced to switch back to Firefox: one example is once there was a build that crashed my browser when I opened a new tab in tree view.
The extension support is in progress as well. UBlock Origin and BitWarden work, but YNAB Toolkit doesn't work too well.
Overall if you are okay with alpha/beta testing a browser it's fine, but if not, stick with Firefox.
I use it as my daily driver every day and I have very few issues with it.
That said I'm a pretty frugal user, I get anxiety if I have more than 15 tabs open, and I shut down every night.
Some websites downright don't work with it, but then you try Safari and it's either the ad blocker or Webkit, rarely Orion itself.
I'm in the camp that likes Webkit rendering the best. I think it's the most appealing font and content rendering out there, even though it feels understaffed/underfunded. I prefer Webkit overall.
I had the same experience. Even if uBlock Origin kind of works, somehow it doesn't block the same number of ads as it does with other browsers, I wonder why. Same uBlock Origin version, same filters enabled, fewer ads blocked on Orion. I used this site for example to test: https://d3ward.github.io/toolz/adblock.html - scores 65% on orion and 99% on other browsers
Yes! You are unique among the 2206539 fingerprints in our entire dataset.
PS: with my vanilla Chrome - I use several browsers, for banking and buying plane tickets I always recommend a vanilla one - I get 47% with many things blocked. Is this thing working correclty? I just checked my hosts file and it is vanilla too. I used to have a large hosts file but this is just inconvenient since it breaks sites. I don't understand why so many things may be blocked in my chrome browser.
> in the end, you can be tracked because you are still unique
Unique isn't a bad thing really. The trick is to regularly be unique in a different way. I'd rather each tab have a unique fingerprint than simply trust that there's nothing about my browser/device that could identify me. There is an ever growing list of data points your browser makes available and a lot of time, money, and research that goes into finding fingerprinting techniques.
I agree it's good to have a locked down browser for most things and other browsers for websites that you can't get working with everything blocked. Brave is worth a try before you hand your browsing history over directly to Google with chrome or to MS with edge.
> Unique isn't a bad thing really. The trick is to regularly be unique in a different way.
This perspective is new to me, and very interesting. If it's intractable to know which parts of the finger could change, the fingerprinting becomes useless.
99% as a score of this reproducible ad blocker test. You can see the individual tests on the site. I'm not saying it's a gold standard or anything, just a tool I happen to use. And you're right, passive tracking is a b.tch :)
> For power users who want that last 10% with uBlock, we recommend disabling Orion's content blocker globally. Running both Orion blocking and uBlock may cause interference with each other.
Orion has actually triggered two kernel panics for me in the past week. It's easy on battery life though, so I've been using it still when I'm on the go.
I'm not very familiar with this space, so perhaps someone could explain this to me. I was under the impression that a kernel panic indicates a problem at the kernel level, not in userland. Is it not the goal of most/all kernels to insulate themselves from userland mistakes?
I actually think it has something to do with webkit and the fact that I use Quartz Debug to disable vsync at all times. I'll try to repro, but anecdotally speaking, my M1 MacBook Pro restarted when I was doing something in the Orion browser and spit out the typical kernel panic "Send to Apple" dialog upon reboot.
It happens on Mac. It's when you get an infinite spinning beachball. Sometimes called beachball of death, but generally called kernel panic on Unix systems.
Most of the time with Macs it starts happening when you have hardware failure. It happens way more on Windows and Linux because of the much larger range of supported devices and drivers and the varying quality of the drivers for them. Most drivers running in Darwin as made by Apple. It's also the reason Microsoft created a certification program for device drivers for Windows.
I definitely had a corrupted CD-R backup that would kernel panic OS X, Windows 2000, and Linux, circa 2003.
Though, most of my OS X kernel panics were due to a GPU slowly going bad and randomly corrupting memory (both under OS X and Linux, when in a particular graphics mode).
QNX is the only OS that I've pushed hard and never seen a kernel panic. BeOS was nice in that my (userspace) ethernet driver would crash overnight most nights, and I'd wake up to a prompt saying "I'm going to restart the crashed driver? Okay?", but I could reliably kernel panic BeOS with some dodgy semaphore code in userspace.
It does have a mobile version for iOS, which is notable on its own because is supports (some) Chrome and Firefox extensions on mobile. It suffers similar beta-y bugs as the desktop version though. Personally I find it worth it to have uBO on my iPhone.
I am hopeful for its development but it’s not a daily driver for me. Too many bugs, not a large enough team to support it. When I have looked at the support forum the developer is sometimes a little hostile about bugs/issues, enough so that I don’t care to report my own issues. This is a subjective statement your experience may differ.
I agree. I love Kagi search but this seems to have no unique features over other browsers. You are better off using your otherwise favourite browser and changing the default search engine as far as I can tell.
It has many unique features over other browsers. To begin with, it is zero telemetry by default. It is also the fastest browser on Mac as measured by Speedometer. And it supports both Chrome and Firefox extensions, in a WebKit based browser. This should be a good start.
Zero telemetry and fastest aren't unique features (assuming that fastest is not notably true, since every browser seems to be the fastest in some benchmark or another). Supporting both Chrome and Firefox extensions may be, but that seems like a very small niche?
Perhaps, but it is the only browser in the known universe that runs Chrome and Firefox web extensions on iOS, which is on billion devices.
Zero telemetry may not be a unique feature to you, but if you care about having a privacy respecting browser it certainly is. If a browser has telemetry, which almost every other browser has, it is not privacy respecting by definition.
Ok, now that is a unique feature! Bringing that large library of extensions to a pretty locked down platform is definitely a selling point.
> Zero telemetry may not be a unique feature to you
Unique does not mean important. It means hard to find elsewhere. There are lots of browsers that have zero telemetry. Adding another one doesn't stand out. It can be an important feature of a browser but it alone won't bring many people to a new browser when they can get it elsewhere in a more established product.
Thank you. This seems to have elicited a clarification from the other user, which I appreciate.
> It means hard to find elsewhere.
Well…"not found elsewhere". It's "the only instance within a given set".
You'll sometimes (perhaps increasingly?) encounter "more unique", "most unique", or similar. These usages dilute the meaning of 'unique' and make it more like 'unusal' or 'rare'. People who use 'unique' this way tend to resort to longer phrases to express 'unique'. Perhaps somebody knows a newer word for expressing 'unique'?
Just to add, the main complaint I have had with it is the memory leaks/bloat. For example, I sometimes listen to music via youtube while working and after a couple hours the video player interface will become unresponsive, you click pause and it takes 5 seconds for the video to pause. This problem has been around for over a year, I remember seeing some similar posts on the support forum but it was kind of dismissive. This kind of bug is a deal breaker for me.
For me the highest stopper is that it does not really work with Google Meet and it is not clear if it is a Webkit bug or not. Apart from that, I could probably use it as a daily driver.
My "favourite" feature about Meet is that it takes me through all these self-satisfied feature announcement popups to explain to me how all these different things work. every. single. call.
I use meet as seldom as possible (basically just when a client, partner org etc sets up a call) but it still adds up to at least 2 or 3 a week at minimum. I wasn't interested in those feature popups the first time I'm definitely not interested now after you've been showing them to me 2 or 3 times a week for the last year or so. There doesn't seem any way to make it stop though.
Careful, someone at Google reading this will green light a new version. Google has gone from talk to hangouts to meet and gotten even more proprietary with each iteration. I’m sure there is someone itching to create yet a new name/version that would have new draconian DRM. Heck, they might even try and bring something like ActiveX back if they knew it would make a users experience worse.
Yes, it's better, but you cannot share a browser tab, only a window or the full desktop.
Edit to add that the suggestion of using Orion for everyday browsing and FF or Chrome/EDge for Google Meet is totally reasonable, but somehow I find quite a lot of friction doing that.
This is how I feel about Vlad's statement on their Discord that he wants to provide email services. That strikes me as a big project in terms of both development and operations, somewhat contrary to their privacy focus, and not something that their users are asking for (on Discord, I'd say that most commenters said they were unlikely to use it any time soon, switching would be too much of a hassle). And while I subscribe to Kagi and love the search product, they're still a tiny company with no long term track record, and far from what I would feel comfortable trusting with my email, which IMO is much more sensitive than my search history. Meanwhile, although Kagi Search already reliably beats Google IMO, there's still a ton of stuff they could do to improve their main product.
I had similar reservations when I read the email discussion. I hope they aren't spreading themselves too thin.
I use Orion as my daily driver (because of the nested tree view), but Kagi Search is more important to me, and I don't want to see that go down under the weight of a bunch of distractions.
Completely agree with you that Kagi e-mail will be unnecessary, even though I trust them on privacy. There are already great e-mail providers, such as Fastmail. But it makes some sense for them to go for e-mail, since they're already hitting Google on search. They're basically going after Google where it's possible: search, web browser, e-mail. They can't go after Google on maps or video streaming, so they do what they can. I'd much rather they innovate and get ready for the future of the internet after AI has destroyed traditional discovery.
The reason people use Google Maps is because businesses provide them with their opening hours and users provide them with reviews, photos, etc. A small company like Kagi can never have that (unless they buy it from Google), so they can't really compete with GMaps. Apple could have a shot, if they weren't so arrogant.
I’m also a paying user (and a fan) of Kagi. I’m still confused by the strategy. If you are not Google and trying to have an hegemony on the web, there’s no obvious link between a browser and a search engine. I don’t see the point from a business perspective. It sends a strange signal to users "Hey, by the way, we are reinventing the wheel and making also a browser that works only on some operating systems".
It’s a bit like going for groceries and then your merchant told you all the time "By the way, I’m selling this car prototype. It makes sense! You have to take your car anyway to comes shopping for groceries so why not buying the one from us?".
I really this analogy. It feels weird especially when Google and Brave both have their own browser and ad programs, so it's in Kagi's favor to stay as far as possible from inventing their own browser.
Second part of the answer to your question is that you can not humanize the web (our stated mission) with just Search. The window to the web is the browser. It is the most important software we have on our computers and yet we are as a civilization somehow Ok with the fact that most mainstream browsers through which we inform and educate are directly or indirectly in the service of advertising companies (!).
If that is not dystopia, I am not sure what is. And that is what we are up against, even if the odds are against us.
Orion is the first browser in 20 years that you can pay for and own. Owning your browser is a thing again thanks to Orion!
It is also a great customer acqusition channel for Kagi, as we do not do advertising and many Kagi members told us this is how they found about the Search product.
Yes it is hard. Yes Orion only exists on Mac, because you have to start somewhere and we are a small team. But we will get there.
100%! There is more than enough to work on in the search engine space without getting distracted by other red oceans.
Is it just that the search engine doesn't make enough space or they are finding it isn't sticky enough, so need to add more services to justify the price? (e.g. email...)
Firefox suffers from the same problem almost all other browsers suffer - it is ad-supported (indirectly in its case). This impedes its ability to be the real user agent, as its customer is the world's biggest advertising agency, not its users.
Like with Kagi search, we wanted to build a browser that gives power back to the user and aligns all incentives. We find this incredibly important.
This is why, unlike Firefox, Orion is both zero telemetry and comes with a built-in ad and tracking blocker. And also unlike Firefox, this makes Orion both privacy respecting and privacy protecting, by default. And that is just a start of things you get when you align incentives in a browser.
In fairness to Mozilla, Thunderbird, Bugzilla, and a few projects like DMOZ were fairly well received projects. Even recent ones like Firefox Send and their email forwarding service are quite nice additions. On the software side, their investments in in-browser translations are money well spent IMO.
It's the stuff like Firefox OS and Pocket gave the bad rep to Mozilla. I don't know what they were smoking about Pocket, but I can see their thought process in making Firefox OS. Chrome OS eventually did become a mainstream OS, it could have been Firefox OS too, if it has the right people and money.
As another paying Kagi user, I feel that this is not a waste of money. A browser that blocks tracking and advertisements by default is not only good for the users privacy and security, it directly hurts their #1 competitor. It's two birds with one stone.
I had similar concerns from the get-go. I honestly don't know how big their team is now but I always only see one or two people post online as representative to Kagi. Won't stop me from paying for Kagi but I definitely will not be paying for the browser. Arc is a much more interesting take on a browser than Orion, and A LOT less bugs.
As far as I know, Orion is the only iOS compatible browser that lets you load extensions from the Chrome store and possibly Firefox as well although I'm not sure about that.
Parent is probably referring to the saga of extension support in FF for Android (it used to support all extensions until around 2019, then it didn't [0], then it did again but only a small number of extensions [1], they only recently got back to supporting a large number, but not all extensions [2])
I highly disagree, google 10 year ago had nsfw content, it had a great long tail. I signed up and found it to have little depth and it skips most of the web. It does a great job of filtering out spam but it equally good at filtering out everything else. Tech related it's great. For shopping not so great. For local not so great. For science, education it's pretty decent. For anything like torrents or coupon sites or sites in other languages it's not the best.
People say this, but it's rarely true. Anybody running a subscription service quickly learns that it's not worth bothering with customers who won't pay at least $5 per month. They're the most trouble for the least income.
I'm impressed by all the work the Kagi team is getting done. I wonder how they split their focus between the two major challenges of building a new search engine and a new browser with the same small team.
Anything proprietary or specific to Kagi needs to be sanitized / made a Plugin. E.g. an error message might say "please contact support at X." It's non-trivial to make all errors generic and pluggable. It's a lot of work to keep the business separate from the OSS part.
I've been using this as a daily driver on iOS for about a month now and it's been great. A few random teething issues but no crashes or real bugs. Supports extensions well.
I think that your usage was actually correct- if I understand correctly, choosing to use Mac's native keychain is kind of "going along with" what Apple prefers- is that the sense of what you were saying? If so then you could definitely say apps "toe the line" when they do this.
macOS has native password autofill which opens the Keychain experience to 3rd party password managers. Unfortunately browser support is limited to Safari (even though the API itself is open), while Strongbox [1] seems to be the only option on password manager side.
Maybe this will improve with the adoption of passkeys, where browsers are actually trying to integrate with the system passkeys API (a neighbor to autofill).
I've been using Firefox's built-in password manager before moving to Bitwarden, and I don't think Firefox's isn't all that bad.
All passwords are E2EE, and the sync server is open source and you can theoretically self host it.
I recently read about how Bitwarden adds Passkeys to Firefox, and was left sour there is no universal API for it. Bitwarden just overrides it as JS level, and hands it over to the browser if there are no passkeys selected/available.
I imagine Safari + keychain is coupled in a more secure and well-defined way.
I'm the opposite, while I don't trust any company, there are levels of trust I give. I trust Apple over a 3rd party. I remember when LastPass was touted by everyone as safe and secure...now look what's happened to them several times.
I mean, we'll see people NOW saying "oh, I never trusted LastPass", but that's BS.
At this time, this is not for me unfortunately as I use 4 different OSes (Mac/Linux/iOS/Windows) but looking forward to checking it out again down the line.
I use it on an old macbook-pro-2012 (stuck on Catalina and on an old version of Safari). Thank you Kagi for providing a build with a recent version of WebKit. On the iPad I use it just for youtube, sometimes. It is still pretty rough around the edges compared to current Safari.
Some bugs still. Mostly if I run into a bug, and I try to open the same site in Safari, I experience the same bug.
Occasionally it'll do something stupid and hang completely, but I'd say maybe that happens once or twice a week, not enough to stop me using it (and I live in it most of the day).
Interesting. I’m thinking of giving it another shot. In theory it sounds like the perfect browser for me. The only extension I use is uBlock and BitWarden so I don’t really care much for widespread support of extensions and it sounds like that’s the last really buggy part.
Currently, I stan Firefox, but only because all other alternatives compromise my privacy and security so egregiously. Mozilla has been messing around for years and in many ways, some merely dumb while others actively diminish the benefits of Firefox over others. I don't think a privacy conscious alternative browser is in any way a waste - it's exciting!
But I'd require both fullsome Linux support and an open source approach to even consider it.
> How will manifest v2 extensions work with Orion after manifest v3 roll out in Chrome?
> Same as before! Orion has its own implementation of the entire web extensions API and different "manifests" are just numbers. We support web extensions APIs regardless of how Google decided to call them or change them. Manifest change impacts mostly Chromium clones, and Orion will support both "manifest v2" and manifest v3" extensions in the future.
--
> Will Orion for iOS also support web extensions?
> Yes, we have preliminary support for some web extensions on iOS.
> Wait, are you sure? No browser on iOS can use Chrome/Firefox extensions!
> Orion makes it possible! We ported hundreds of web extension APIs to run on top of WebKit, which also runs on iOS.
Thanks, not having an iOS device or Mac I didn't check that far into their site. I'm not willing to invest into that ecosystem until Apple allows some freedom.
Depends on what platform you use - iOS is quite buggy whereas I have close to zero issues on Mac. Turn on compatability mode (left button next to address bar) if a site does not work properly which turns of ad blocker etc.
I tried daily driving Orion for a couple weeks and there's a lot of great work being done, but I had some pain points. The extension functionality on iPadOS is really jank in a way I don't quite know how to explain. Like sometimes I come back to Orion and the extensions are signed out of or sometimes they straight up just don't work. I've also had some bad one off experiences of random sites taking a long time to load or running into issues. Orion plays really well with Safari though, which is great, so I run a very strange combo of Orion on my iPhone, Safari on my iPad, and Arc on my Mac. There's enough additions on top of Safari for me to prefer Orion over it (if it works well), but not enough for me to prefer it over Arc on my Mac. Arc's QOL features and out the box UI design just put it way above any browser I've used.
Not sure what you mean with "lighter", but the reason some browsers feel heavier is the security features that introduce a performance penalty or take up more resources, like jitless mode and process isolation for example. Firefox is not a secure web browser.
Funny that it uses Hacker News as its example rendering page. We must really be target group. It's also not a particularly hard to render page or one that tracks you.
Looks of the browser are pretty great and it has built-in most the features you would need.
Some extensions do not work. E.g. BitWarden autofill does not work either as Chrome or Firefox extension.
Some websites get laggy. Sometimes ChatGPT page lags so much that is unusable. Also if you browse GitHub, some views will lag very much, and make it unusable.
Maybe some day it will be competitive also in compatibility.
But since it is MacOS only, I always will need to use many browsers...
Yes. But maybe Reddit is not the best benchmarck due to their user-hostile behaviour anyway (well, maybe it is "best" because of this). The only purpose of mobile website is to make the user to download their app.
This is interesting... I used to daily drive Safari but the kagi extension somehow bricked the entire Safari. I've tried all the tricks to clear out ~/Library folders but it is still broken.
Back on Firefox now... but this could be worth a spin.
Is it ever gonna be possible to support Safari Extensions and Time reporting/blocks via iOS?
I can set time limit for HN if i use an app, or if I use Safari. Switching to Firefox (with the same WebKit backend) means no screen time limits apply.
I'm not using macOS or this browser, but my guess, seeing how the website mentions it's native app, it probably has to do with ibtegrations with certain features available in each macOS version. Live text, sharing options, etc.
To be fair, Brave seems to have genuinely tried, but, seems to be failing to reach sustainablility judging by their various pivots. They've tried to walk the line of leaning toward the user while middlemanning revenue ideas to said user. BAT and crypto, VPN, messaging sales, now search+ads. The market seems to be saying no thanks from what I can see from the outside.
Orion, in straight up exchange for cash via donation and Kagi subsidization, offers to be entirely the user's agent.
Gets stuck on major websites. The extension support sounds extremely fishy: I could install uBlock Origin, but either they are violating the Apple AppStore rules, or the extensions aren't actually working (fake API, not really implemented).
You cannot ship a JavaScript runtime in the AppStore, so you cannot ship a browser that is not a safari WebView, so extensions like uBlock Origin cannot possibly run on an iOS browser.
So which is it? They cannot get away with violating the appstore terms. It must be a shim API that marks the extension as installed to some degree.
Developing a complex product like a browser without telemetry will be hard. Especially if they are trying to commit to objectives like using minimal battery or memory. There are many many different websites so its hard to predict what your users are using the browser for.
I'm glad someone is trying to doing something hard that will benefit people. We need more companies that want to do the hard thing, not the easy profitable thing.
Telemetry would allow them to make their product better for more people and know where to invest to benefit their users more. In this case proitability is proxy for how much value people get from the proxy. The thought that something is bad because it is profitable is misguided.
Telemetry is a cheaper solution to hiring quality assurance. Many will feel sharing their usage details to increase profits is acceptance and others might feel this free product isn't so free anymore.
No, collecting information on how many times someone types the sequence "abcd" is a waste of resources. Figuring out what telemetry will be useful and understanding the privacy implications is important.
>Telemetry is a cheaper solution to hiring quality assurance.
These are not comparable. Telemetry is about understanding how the product is operating. QA is about testing the product.
The main downside is the extra work and cost of getting it all set up, properly cleaning up unneeded telemetry, and making sure telemetry itself doesn't cause performance issues.
>No concept of misleading info or bad product development direction from using it?
Telemetry doesn't replace a strong protect sense, but without telemetry you are flying blind.
Telemetry can be privacy respecting. For example it may be useful to know the average page load time. The browser could record 100 different page loads and then send the median time.
It almost can not by definition, as every telemetry will include user IP address which is considered PII. Regardless what the stated purpose of that telemetry was, private information will leak and that browser can not be calling itself privacy respecting. Zero telemetry by default is the only way a browser respects user privacy.
If the user opts-in into telemetry then obviously it is user's choice, but usually browsers with telemetry do not make this an opt-in choice.
The extension support is in progress as well. UBlock Origin and BitWarden work, but YNAB Toolkit doesn't work too well.
Overall if you are okay with alpha/beta testing a browser it's fine, but if not, stick with Firefox.