> Shared CEF runtime across apps. Every app currently bundles its own CEF copy. A managed shared runtime would drop binary sizes to a few MB per app. On the roadmap.
This[0] sounds interesting. I am not familiar with CEF, so I wonder how the versioning works. When different apps require different versions of CEF, do we just essentially end up with the electron model where every app bundles their own browser (just slightly less bad). Or is there still an advantage to a "shared runtime" in that case?
The biggest weakness of a framework like Tauri is the choice to target system webviews instead of bundling a browser runtime.
It seems great to be able to cut hundreds of megabytes out of your app installer, but the platform differences wind up being a complete and ongoing pain in the ass.
Tauri support on Windows is phenomenal.
Tauri on Mac runs into lots of WebKit/Safari issues, especially on older Mac machines that have an older engine that doesn't support modern web APIs. Your app can crash or be left non-functional. You'll find out about these runtime bugs in the wild randomly, and patching for some customers can take days, if not weeks.
Linux support is hellish, and it's best to not even try targeting Linux with Tauri.
Tauri is in the process of adding CEF support. It should probably become the default build target for all platforms.
I use Wails which is Tauri but for Go and I don't have the kind of issues you're mentioning. Maybe that is a difference between Wails and Tauri but I don't think the system WebView is a significant factor.
Are any of your Mac users using an 10-year old WebView? We frequently ran into that. And there's nothing that can be done about it except engineering around it.
I also doubt it works well on Linux. The performance of webkitgtk is like running an emulator inside an emulator.
Regular Tauri app (aptakube) user on linux here: the experience is very adequate and smooth, I have no complaints.
Speed benefits relative to Electron (similar app: K8S Lens) alone are enough to deal with many possible issues.
Could be attributed to app developers going the extra mile, but I suspect it's the framework choice.
This point of view always confuses me, because web developers already need to deal with platform differences. Especially if your app app also runs in a browser, like Slack and Discord — at that point, what issues do the differing system webviews cause that you don't need to deal with anyway just targeting browsers?
It's also funny to me as someone who's been building websites for 20+ years at this point, because the platform differences used to be much, much worse. Coincidentally, I just saw this article, which makes the case nicely: https://www.bram.us/2026/06/21/do-websites-need-to-function-...
Not necessarily? As far as I know they come from the same WebKit source. But the bridge between WebKit and the OS is of course specific to each system. WebRTC is still experimental in webkitgtk while caniuse says Safari supported it back in 2017.
I remember encountering one bug with Final Form that triggered rarely in Safari, 100% in Webkitgtk, and never in Chromium.
Here's the developer of Tauri saying it's hard to recommend webkitgtk / Linux support:
> So if you need good linux support now/soon i can't 100% recommend tauri (for Linux) as of now. (I used to be more "forgiving" but with webkitgtk getting worse/more unstable each release i changed my mind)
Sure, but many people use Safari, which runs that exact same WebKit engine under the hood. So if your app is available in the browser in addition to Tauri, you have to support it anyway. And at the very least, you as a web developer should be used to supporting it.
Many people run older versions of Mac OS X (because their machine won't upgrade), so their safari is aging -- they use an alternative browser, firefox and chrome support older mac os x.
Of course using unsupported OSes isn't the best idea, but Apple give you security updates longer than they give you 'fun' safari updates.
The platform webviews are significantly older/worse than typical web browser versions, especially on macOS and Linux.
On macOS, the only way to upgrade your WebView is to upgrade your OS, which requires rebooting. Lots of people just don't bother. You can upgrade to the latest Chrome or Firefox just by downloading it (assuming they support your macOS version), and they auto-upgrade themselves pretty aggressively.
On the web, very old versions of Safari (6+ years old) are a tiny fraction of a percent of your traffic; many web developers just ignore them. In a desktop app based on a WebView, ancient WebViews can be as high as 10-15% of your macOS user base. Ignoring them is not an option.
On Linux, it's common for the major version of WebKitGTK to not upgrade at all except during major OS upgrades. Anyone on Ubuntu LTS 20 is going to have a 2020-vintage WebKitGTK with security patches. (And Ubuntu LTS 20's WebKitGTK was buggier than macOS WebKit, even in 2020, because Apple has more dedicated full-time developers and testers making sure that macOS WebViews work end-to-end.) If you're shipping an app based on WebKitGTK, you can expect to see double-digit percentages of your Linux users running really old WebKitGTK.
Maybe you're such a great developer that your web app works great on ancient browsers, but, if so, it's probably because you didn't need/use much JS in the first place. (Maybe you used HTMX or something.) In that case, is there even any benefit in shipping a "desktop app"? What's your desktop app even for? Offline support? (But your app is all server side…?)
If you have a JS-intensive app that works great on ancient, buggy browsers, then platform WebView might work for you. It's not nobody, but it's hardly anybody.
Yes, we're used to platform differences, but that's kind of the benefit of shipping a desktop app, isn't it? So that we don't have to deal with the platform differences. If you bundle the browser, then that ought to all go away. No graceful degradation necessary, no polyfills.
Tauri's cef-rs is indeed quite good. I think it's available, ready, works: it's just that there aren't many folks using Tauri and only a fraction of them are aware/interested in exploring further. https://github.com/tauri-apps/cef-rs
Huge fan, this should definitely be the default. The user experience is incomparable.
One thing I'd like to verify: can the OS effectively use shared mem for cef across multiple different cef-rs apps? I really hope so. In this time of RAM being scarce, this optimization could be such a benefit.
Or use QtWebEngine which uses Chromium under the hood, provides a stable and rich API and enables sharing rather then bundling the browser runtime + is Cross-Platform + security updates are handled too.
I love how we’re now reinventing the browser as a much worse version of itself. What if instead of one or two general Web browsers we make everyone install 10 random versions that can only open one website?
True for this decade, but in the previous decade it was very much the opposite. Before you used any kind of browser api or nice language feature you would feature-detect it:
if (typeof Array.prototype.includes === ‘undefined’) { …
And if it wasn’t there you would define it yourself, it was called “polyfilling”. This was so commonplace that we built significant tooling like babel to standardize feature detection tests and fallback implementations - for a few years you could write
request.then(response => response.json())
And behind the scenes the Rube Goldberg machine would turn this into something that would run in a JavaScript environment that had neither arrow functions nor promises.
I doubt the benefit. Practically every Electron app on a desktop uses different versions of Chromium and many are very out of date because of the risk of breaking when upgrading.
People build web apps for an array of browsers and huge ranges of versions. I think if you started using some tech to deploy an end user program and knew from the beginning the browser could be updated beneath you it would work just fine. But if you start with a golden version of Chrome and put off updating for too long you’ve let yourself get too comfortable.
De facto they do because functionality built three years ago and tested then is running along side functionality they built yesterday and tested on today’s Chrome.
People also do seem to test on iOS Safari because that pain in my ass needs special care on my software. So if a site works on it they either got lucky or tested on an iPhone. It’s generally only other people’s weird tech demo stuff that doesn’t work.
I agree and disagree, you can't target everything, but most (not shit) devs will target at least Safari - 1 or 2, simply because the iPhone market is too good to miss out on. And Safari being, well, Safari, means targeting that is a pretty safe bet for anything else.
Depends on the region, no one where I work has an iPhone or a current Mac, so stuff gets tested on FF and Chrome, and Safari gets thoughts and prayers. We would test on Safari if it were simple, but alas.
Maybe for your app, it doesn’t make sense. And if it’s a pure enterprise app, fair enough (assuming it’s an enterprise that was started more than 15 years ago and only targets regulated or very specific markets). But a good way to guarantee that your app will never go beyond Windows desktop users is to ignore the most dominant mobile platform by users who actually pay for software.
We are a very small company, and have always had far more Firefox than Safari users. And though they get by via dominance, IE style bundling of the browser to the OS is toxic, so good riddance.
You could, but by targeting a specific Electron app the mindset would be much simpler. Just take a look of how many times does the dev behind VS Code decide to upgrade their Electron/Node.js version, and how many breakages due to them.
Just to let you know, CEF was used for Riot and League of Legends client as well [0]. The results haven't been nice, but I'm not aware if this was a problem with the CEF technology itself or other component/processes are to be blamed.
When the new client was built, microservices were the hot new buzzword.
The new client is some weird plugins/services based architecture. Things that'd barely warrant their own class in a boring OOP-based UI framework are instead now "isolated" services. The reality of this isolation being that if one piece breaks, the whole UI becomes unusable anyways. Dozens of things that in another app would've been just a simple synchronous call now behave like remote procedure calls and messages, forcing all the complexity of distributed systems into a local application for no reason.
That's why it runs like ass, breaks if you look at it wrong, and your CPU draws more power when using the client than when playing the game at 200FPS.
This is more a function of how mismanaged the project was at Riot (the iron client days, choosing Ember, etc.) which led to the current state of the launcher.
I regularly install and uninstall the league client on Mac based on how I play that game (require fresh installs to raise the playing cost) and their client really sucks. Even freshly installing it, the client inhibits and hijacks mouse clicks from the full screen game when it's open. It took me months to figure out I would have to minimize the game, open and minimize the client (separate app) open, and then clicks would sometimes properly return.
Before that I was closing both and playing a game of roulette. By the the time my game was working, I may have already been reported for being afk and the game ended. Edit: other bugs have included starting to download the game and then signing in afterwards (their UI doesn't stop you) and then download progress disappearing, perhaps hanging, and you having no way to know it for a 40gb file unless over an hour has passed and you check disk size and then realize the client doesn't know how to load it. Start over and do a fresh install again, clear cache etc because their cache of the client still thinks somethings being downloaded even though it's not. Also having chat permanently off, results in weird glitches with friends requests and being unable to add new friends.
Horrible experience. But since the game is so optimized for addiction and dark patterns these days, and sunk cost, its a game I find myself returning to every once in a while.
Both Steam and Battle.net use CEF for their UI as well. And IMO they are on 2 ends of the "nice to use" from the implementation side (Steam being a sluggish hell and B.net being nice). Though then again B.net is only for Blizzard games, so they can also optimise for the limited set games.
I'm so sincerely glad that 28 years later I get to spot hate for eMachine in the wild. What an absolute piece of shit. I remember buying one and immediately returning it and to this day... you get it.
> Because we can’t differentiate between a legitimate sign in and a MITM attack on these platforms, we will be blocking sign-ins from embedded browser frameworks starting in June
Granted this was years ago, maybe the situation improved? I had to abandon my CEF project because of this.
Most apps (on desktop or mobile) open third party auth flows inside the user's default browser, which makes this a non-issue. For one, if you embed the Google login flow into your app then I can't reuse my existing session in my browser. But it also exposes my full credentials to your app for no reason, which is a good thing to avoid.
I was wondering how this integrates with Deno's permission system, which is one of its biggest strengths especially for letting agents run amok on your device.
The CLI reference page[0] notes,
> The permissions you grant at compile time are baked into the compiled binary:
I think it would be nice if this could be surfaced to the user somehow, like letting the user know and decide which permissions they want to give access to.
> Runtime permissions for desktop apps (a permission prompt on every filesystem / network access, i.e. Deno's permission system applied to desktop sandboxing).
You are running a binary that you got from the developer. If it presented you with Deno permissions, I think that would be misleading because there’s no guarantee of their integrity.
That is true. I wonder if it could be possible to let the user supply and wrap the app around their own, trusted installation of Deno (rather than the one bundled in the app) to specify permissions.
Then do `deno ./my_downloaded_deno_gui` instead. If you trust the copy of deno you downloaded, then hopefully it can be trusted to verify the permissions of random downloaded deno-apps.
Yes, it kind of defeats the standalone binary aspect but if you're really concerned about security, maybe it's a happy medium.
Deno continues to impress me. It’s honestly been quite a while since I started a new project without it. It has fully won my support over Node.js, the ecosystem has really matured nicely. I don’t know how often I’ll use this feature, but it’s really nice to have the option!
"NAPI allows native Node dependencies to target Node without actually depending on a specific version of Node or V8. NAPI abstracted the V8 JS engine away, by providing JS-like APIs for: creating an object, declaring a property, etc.
NAPI is the contract that all modern native Node modules use to interact with Javascript." [1]
I'm still using node/npm and it's... fine. Every so often I read these posts and think I should change but node/npm is already a low friction part of my workflow.
However, what I have seen is that a lot of other libraries I use have switched to Bun. I haven't seen any that switched to Deno, and so I've been under the impression that Bun is becoming a strong node replacement candidate while Deno is not, or at least that the community is making a strong preference for Bun. Anyone have more insight into this?
I'm happy to see this I see that this provides CEF, Webview and Raw * backbends but it would be nice if there was also a launch in browser option (like WebUI has). To me that has the best tradeoffs if you want to avoid the mess that is webkitgtk but still not ship (and be in charge of updating) a chromium engine with your app.
> I'm happy to see this I see that this provides CEF, Webview and Raw
They beat Tauri at their CEF support.
Webviews are a mistake in most cases. They're too platform-specific, and certain Webviews (Safari/Webkit) are buggy as hell, making platform support a nightmare. (Linux, ironically, is even worse due to how underbaked webviews are on the major desktop Linuces - Tauri is barely functional on Linux.)
Deno Desktop could be a real contender in this space. It's good to see more Electron alternatives.
I think that my Sciter is better option when you need HTML/CSS/JS native application running on
Windows (XP and beyond), MacOS and Linuxes.
Sciter SDK [1] contains scapp[.exe] - standalone Sciter engine that can be attached to HTML/CSS/JS bundle making standalone (single exe file) and portable executable. https://quark.sciter.com/ tool allows to compile such apps.
Size of "hello world" is a size of scapp.exe binary + size of compressed HTML/CSS/JS bundle.
On Windows scapp.exe is of ~14 Mb. On Linux ~18 Mb.
Linux version at startup detects GTK4, Wayland or X11 and uses those as windowing backends.
On all platforms Sciter provides out of the box: HTML/CSS/JS runtime, libuv based Node.JS alike runtime, GPU accelerated rendering, WebGL 3D runtime, JS built-in persistence (NoSQL DB).
It does not have TS compiler built-in as Deno, but that TS-to-JS compiler is better to be outside anyway as it is used only once - at app loading.
Wouldn't that just be "Raw"? I.e. start a webserver and ask the system to open the URL. There is no "special stuff" to do in this case like avoid sockets in favor of IPC to a well known webview or package CEF and no real integration to make with dev tools etc after - it's just open socket and serve from prebuilt binary.
No. As I understand it, the Raw backed just gives you a Window with input handling and you have to embed something like Skia, WebGPU for the graphics. So basically you have make your widget library yourself.
Now you can just start a server with deno pretty easily and serve a website. But WebUI will actually also manage opening the browser window for you as well a make the communication between backend and frontend just like using a Webview or electron.
Raw still exposes access to Deno APIs & ability to call native code but it doesn't seem to assume it should give you a window out of the box or anything based on my testing (AFAICT you have to orchestrate most all of that yourself, Deno just won't get in the way with packaging other stuff like CEF automatically). If this is just some issue on my machine with the canary build though, the existing "Deno compile" should accomplish the same goal for a binary version without any GUI components at all.
Outside making it so you don't have to call to open the link yourself, I'm still not sure what could be integrated in the scenario. Deno can integrate with WebView because the WebView APIs were designed to allow external applications to have full control of the session. CEF (the Electron-like) approach works because Deno packages CEF & the APIs as part of the app itself, having even more control of everything. Browsers are meant to be in control of themselves or the user, and have had a long history of fighting malware trying to act otherwise.
> Bindings are not IPC. The Deno runtime and the rendering backend run as threads / processes inside the same address space (CEF) or coordinated process group (WebView). Calls go through in-process channels, and the backend dispatches them from its run loop. -- https://docs.deno.com/runtime/desktop/bindings/
I don't understand how the coordinated process group works. Doesn't that mean in this multi-process mode it must be IPC? Maybe the claim "shared memory space" is more an architectural description than an OS-level claim?
My guess is that it's not using IPC on OS level, like D-Bus on Linux, but rather a supervisior process starts and orchestrates child processes as needed. And all these processes use a shared memory model.
from what i understand after a quick look at the source is it uses a C ABI to communicate between the WebView/CEF "host" application and the deno runtime which is loaded by the host as a shared library.
marshalling of values back and forth between the JS/C++/Rust layers still has to happen but these are just straight C api calls in process under the hood so much less overhead than having to do serdes across a socket/pipe.
I'm happy for competition in this space, specially because Deno can run true TypeScript directly and not just strip types like the current Node implementation.
With that said, this is going to eat a lot of Tauri market. Why would I use Tauri now? The 150mb of additional bundle size is just an extra 1 to 10 seconds of download time in most internet connections and you get a reliable rendering engine.
E.g. Tauri uses WebKitGTK on Linux, which has historically been slow, unstable, and frequently lagging behind the main WebKit project. This is enough of an issue that even Tauri is working on the ability to use CEF instead of the system web view in Tauri apps.
Things are generally fine on recent versions of Windows and macOS. The system web views on these platforms will be evergreen versions of WebKit or Blink. But if you want to support very old versions of Windows or macOS, you might choose to use CEF instead of wrestling with Safari-from-five-years-ago.
Deno also just strips the type annotations when running TS code - at least by default. To get type checking you'll need to run via `deno run --check`, or use the separate `deno check` subcommand. No big deal since type checking and linting usually happens automatically in the IDE during development.
Only for non-erasable Typescript! The support for running erasable TS is here to stay. It's not ideal, but on the other hand this is paving the way for the TC39 proposal - if that ever progresses...
Tauri doesn't lock you in to one JS ecosystem. In fact, it doesn't require you to use javascript at all.
Also, we've had several developer framework startups get acquired -- Astro, Nuxt, UV, Bun, Vite. It doesn't exactly inspire confidence in a software that you want to last and give support for years.
I don’t understand this logic. The acquisition doesn’t guarantee the project dies and staying unacquired doesn’t guarantee the project continues.
The state of OSS funding is precarious. The acquisitions at least guarantee some runway for the maintainers. Maybe the acquirer has alternative intentions than to simply bankroll the project’s the way it was, but at least someone is paying to keep the maintainers fed.
After having worked in the JavaScript ecosystem for over a decade, I don’t think _any_ project of significant size is guaranteed to last or have support. We need to be thinking about resilience (_when_ the framework stops being supported), not pretending like using OSS projects without paying them should magically give you some support contract/assurances.
> Deno can run true TypeScript directly and not just strip types
What exactly do you mean by that? Because no js engine carries the ts types into the runtime as far as I know. Deno and nodejs both use v8 as the runtime. v8's internal types are not connected at all to the ts types regardless of the wrapper. The only difference might be when/if static type checking is performed.
I assumed they meant more than that because nodejs does the same since v22.18 and prior would do it with a flag. But they mentioned type stripping, which is really all transpiling of ts does for most code.
Temporarily at a place with 10-15 mbps. 150 MB is around 1 minute.
I grew up on 30 mbps. >= 100 is all I need nowadays.
But 10 mbps and websites and downloads really start to take a while.
The more bits and bytes you save, the more people will be pleased with your stuff! Even if they don’t know what bits and bytes are, and just go based on impatience
>I'm happy for competition in this space, specially because Deno can run true TypeScript directly and not just strip types like the current Node implementation.
this is misleading. there is no "running true ts". you will always be running pure js (until someone actually develops a "true" ts engine), and deno does "type stripping" just the same. the only difference is that it bundles the tools and makes it transparent and config-free which is more convenient (although more rigid).
The same set of reasons anyone made desktop apps before Deno desktop launched. Deno desktop is not the reason desktop apps are made. If someone really wanted a desktop app made before, deno desktop isn't going to suddenly make new ones start appearing. It will just be a new choice among the many.
The overall feature seems really solid, but I'm impressed they couldn't reduce the average package size further from 40MB even when not using CEF. I guess that wasn't a huge focus when developing this feature? Tauri and Dioxus can easily hit less than 5MB for package sizes.
I find the feature matrix comparison to be extremely well done and the sections beneath explaining advantages and disadvantages to be some of the best docs I've read recently.
Deno Desktop is bundling the V8 JavaScript runtime so it can have JavaScript on the backend. Tauri uses rust for the backend and your browser's JavaScript engine for the frontend.
> Web technology is the most widely-known UI toolkit in the world.
Poor choice of words there IMHO.
The reason Electron apps get a lot of flak is because they are everything _but_ a UI toolkit. They consistently miss the mark in adopting UI patterns from their host OS.
Web tech is just web tech. Yes it will allow you to render a button, but even unstyled, the button won't necessarily look native to the OS, and will vary between browsers.
I have seen users having trouble with pixel soup UIs. They may not think "This should be in a native toolkit", but they do think "How the hell do I subscribe to a folder in the new Outlook?".
Right, but bad UI's was not uncommon before webviews, if anything the spartan-ness of the web often simpified patterns whilst reliance on weird hotkeys in desktop apps isn't uncommon.
>reliance on weird hotkeys in desktop apps isn't uncommon
The only examples I can think of are actions that are intentionally not easy to reach. How exactly it's done is platform-specific, but the (mis)feature doesn't come from the platform and can be implemented in other ways on other platforms.
- Apple UIs hide some power user functionality behind obscure hotkeys
- Similar: Shift-Delete to permanently delete (Windows, KDE)
- Similar: Shift-right click to "Open With..." (Windows, KDE)
- In desktop apps that FOR SOME REASON try to look more like web apps, the hidden menu bar can be restored with Alt or Alt-M (Firefox, KDE)
The problem in these usability cases is pretty much always layout and constant redesigns rather than the exact theme the button has. I've seen plenty of unusable native ui soup UIs and very clean and simple custom UIs.
You could call it the exact theme when a clickable UI element looks just like regular text (it was not inside any kind of button-like shape in the Outlook case that I saw), but it's super common to have that in web-based UIs.
> ironically the only group of users I've found that actually care about native UI, are other developers and Mac purist.
One group of people who routinely carry the can for poor product usability and another who by definition care about the Mac platform; entirely what would be expected.
That’s why HN users constantly advocate for Vim, a program in which every single thing works completely differently from every other modern application.
Yes, if there's one lesson from historical UI research that still holds, it's that mode switching is expensive. That's why people install vi plugins everywhere.
Consistent like what? Like maybe a decade ago one could say that osx was consistent, but nowadays even SwiftUI and cocoa is visibly different, let alone every second app that uses electron. And people don't care.
Windows has like 4 frameworks available on a bare new, latest OS install, just go deep enough in the "settings" or whatever they call it, and you can reach down to winforms. And on top the start menu is a react element!
(And in Linux you have the gtk and the qt world, and everything else)
Within OS consistency is much less of thing a thing than Web design conventions. Windows by itself has had several different UI frameworks over the years, so different "native" Windows programs can look completely different from each other.
> Within OS consistency is much less of thing a thing than Web design conventions.
Sorry, are you saying that two random web apps will typically share more UI consistency than two random Windows apps? Because, although I'm not currently familiar with Windows, I would be amazed if that were true.
Great for the developer. The user doesn’t use Mac, Windows and Linux. Just one for work and one at home, with mostly different apps, so they couldn’t care less if it looks the same on different platforms.
They may care, however, if they get anything at all. I do not have the resources to target something to all platforms, so the alternative wouldn't be "Users get UI targeted towards their OS", the alternative would be "Users get nothing since developers don't have the time to also target their system".
Time. I have the time to maintain a single GUI. I do not have the time to maintain three GUIs. Of course they'll be 99% the same, but checking that this 1% difference behaves and looks fine accross systems adds a substantial amount of effort that I simply can not support. And for what exacatly? I want them to be identical accross systems, not different.
A foolish consistency with terribly designed shallow superficial desktop user interfaces dreamed up by overpaid cocaine addled corporate boutique brand designers with not only no experience but actual burning contempt for usability and human factors and accessability and affordances is the hobgoblin of little minds.
That doesn't say anything about the value of whatever UI kit is in place, being shared consistently by apps. A virtue that, apparent from this thread, is no longer universally shared.
OS-level consistency is also consistency. It depends what we value. A lot of apps’ design could’ve been basic, OS-like UI. Apps such as GymBook or WhatsApp are internally consistent while still adopting many elements from the system’s design, instead of reinventing the wheel.
There are two types of consistency in this context: consistency within an OS and consistency across them. I, too, prefer the first because I only really use one OS, but this preference varies. I don't think it's right to say that the first case = "ui toolkit", but the second case doesn't.
How is it a poor choice of words? It might not be "native" UI, but they never claimed as such.
I've always felt that native UI on Linux always looks incredibly ugly and I'd much rather use a nicely styled HTML+CSS layout instead.
In my experience, Electron mostly gets flak for being bloated and slow, it not being native is sometimes a secondary point people add on top.
I've always wanted to build a direct-browser integration that could use HTML+CSS for the layout, but avoids needing a JS runtime. Idk how lightweight servo is but one day I hope I will see my idea come to light
I don't mind the idea of using HTML components and widgets to style desktop applications. CSS and the DOM are widely known and reusing those for desktop apps is probably a good idea.
The problem, as you've pointed out, is that electron apps are bloated and slow. If they became the default and my editor, chat client, terminal, and everything else that I keep open were just thin layers around web applications, I'd rather figure out a way to move things into a browser rather than pull a piece of the browser out to wrap these applications.
There’s a point where it stops scaling in the browser, whether it’s due to scale or poor practices. For example Slack is annoyingly slow to start up and work in my FF, but works OK as an Electron app.
> I've always wanted to build a direct-browser integration that could use HTML+CSS for the layout, but avoids needing a JS runtime. Idk how lightweight servo is but one day I hope I will see my idea come to light
Blitz (https://github.com/DioxusLabs/blitz) is exactly that. It's a new custom browser engine supporting standard HTML/CSS with a native Rust API (and optional integration with Dioxus which is a React-like UI framework in Rust). Baseline binary sizes are around 10mb.
We share a few components with Servo (Stylo the CSS engine (also shared with Firefox), and html5ever the HTML parser), but we've built a bunch ourselves too: notably we have our own layout engine, DOM tree and event handling. Servo is unfortunately tightly coupled with SpiderMonkey, and there is little prospect of removing that dependency in the short term.
That's really nice. It would be really good for game GUIs too where the situation is quite poor and would work well with underlays/overlays/worldspace UIs. That said while binary size may be around 10mb, it still baloons to 500mb at runtime for your TODO list example which is more than some electron apps.
Yeah, the RAM usage is quite high atm. The good news is that it's almost all the rendering layer (WGPU+Vello). So if we can make the renderer more efficient then it's likely we can bring that down. There is also some low hanging fruit in the DOM implementation (but I think that's only actually causing a few 10s of MB of usage).
I would also note that native toolkits (SwiftUI, etc) tend to also use at least ~100mb RAM these days. A lot of that is unavoidable if the app is actually visible, due to modern screen resolutions being so high.
Every time I use Zed across Linux, macOS and Windows , I'm amazed stable and performant it's GPUI framework is. As a user, I'm very happy with it; of course some basic features like accessibility is missing but I'm sure it will be implemented soon.
As a developer, I'm not sure what's the barrier for entry is apart from Rust then again it's the USP as well.
With Dioxus, program logic compiles to native code instead of running it through a JS engine, and it ships its own HTML renderer (Blitz) instead of bundling a whole browser. So it's a lot more lightweight and performant than Electron.
As a minor bonus, the live-reload is also faster than what frameworks like React do. It truly has subsecond latency, which isn't exactly a game changer but is nice when iterating on visual details of an app.
For me it's not stable. After they change their renderer from one to another it freezes for me from time to time. But on the other hand I'm running Nvidia on Wayland so I feel no hate towards Zed owners - and restart is super quick ;)
Interesting, I'm on Nvidia in Wayland most of the time too and haven't had single freeze or crash in a very long time. Even the Windows is running inside the Qemu.
> basic features like accessibility is missing but I'm sure it will be implemented soon.
Accessibility implementations frequently are more complex than the entire UI drawing bits. Most custom UI toolkits never implement accessibility, even decades after creation.
That's good news, but I would continue the timer there. That issue is from March 2023 so the counter is at 3 years, 3 months. Let's see if they have anything decent before 10 years.
Even in a "post-vibe code" era I wouldn't want to create multiple versions of the same app, and none of the "platform-native" GUI toolkits run on everything.
SwiftUI is apple-only, gtk has pretty bad compatibility on non-linux, qt is decent but requires C++ or python, and even so still not much for mobile. Don't even get started on "Windows frameworks", because as I write this sentence they may have left a new one in the ditch.
Flutter may be the closest, but why didn't they go with e.g. Java instead of a new language?
So yeah, if you want a truly universal UI then web is your best bet.
Right. If you want your app to look the same, custom way, ditching what the OS has to offer.
Some developers still believe an operating system has useful UI components and patterns worth adopting. From this thread it's clear that there's plenty who don't. Personally I view that as a regression.
Well, maybe Java's AWT has been correct all this time.
Of course there is value in having "OS-native" buttons, transitions, windows etc. And many parts of GUIs are basically standardized. The problem is all the parts that are not, and have to look the same everywhere.
The OS-component-oriented approach was defensible in the days of desktop only (though personally I think it was a mistake even then), but it's not sustainable now. People want your app on their PC to look and behave like your app on their phone (whichever combination of PC and phone they happen to have), and that's a lot more important to them than having it look and behave like other apps on their PC.
> People want your app on their PC to look and behave like your app on their phone (whichever combination of PC and phone they happen to have), and that's a lot more important to them than having it look and behave like other apps on their PC.
They don't. No one generates GBs of simulation data and create plots on their phones, and this is what "app" I am working on is doing.
Well, getting a hardware-accelerated blank buffer onto a screen to render video content is hardly the epitome of graphical user interfaces. VLC has very few and basic UI elements.
Jetbrains is a better example, using Java with Swing which is not a common choice. As seen from my other comment, I do think this is a good direction, but it ain't any more native than Flutter or for what it's worth an Electron app, none of these are "what the OS provide". FYI Jetbrains has to do quite a few platform-specific tweaks to make them better citizens on each platform.
Portable applications is not a recent need. The only requirement is to have a standardize interface and an implementation for each of the platform. Where you put that interface is an engineering skill.
VLC went with QT (which has done all the hard work) for the UI, and their own libraries for the media playing part. Other software like Emacs and sublime has implemented their own UI libraries. Some just ship libraries and others build UI for them.
The issue with Electron is that it brings a whole jungle and a gorilla holding the single banana that the devs actually need. And the dev flung the whole thing at the users. It's like establishing an iron mine, a steel factory and then pollutes the whole region when building a quick stone bridge would do. Because the only thing you know are suspension bridges.
VLC is not the example you're looking for. Written in Qt for desktop and their own libVLC wrapper for mobile. Yes, in the case of VLC, parent comment is right: you clearly see the issue. And a media player is a relatively uncomplicated piece of software, UI wise.
I never thought I'd defend Electron, but I'd rather use the bloated web UI than a vibe coded Qt/GTK version I'm positive will not have seen any human QA.
I have better things to do than spend my time adopting UI for various different systems. If Electron gives me the option to easily create a UI that looks the same everywhere, then I'll pick Electron over anything else any day.
Nah they stopped caring as well. Developing an application for macOS is hell nowadays. I hate the state of things but both Apple and Microsoft dropped the ball. Linux is even worse, so yeah I don't see a reality out of this unless we actually create something that surpasses the web in all measures.
Depends on the tool. We (mac people) tend to prefer native toolbars and settings menus, but I would say the days of relying on a "native" textarea or button are now behind us.
The other thing I find most Mac people appreciate is a shared understanding of hotkeys and if your app goes against the norm, one of the first feature requests will be to add configurable hotkey support.
Unfortunately, Apple has dropped the ball with their newest native apps in regards to UX and it will take years for them to go back and improve things. The new OS this fall is aiming to start that process, but it will still be a band-aid in some respect.
That is not why people use Electron. The goal is not and never was to just be a "UI toolkit" and "adopting UI patterns from their host OS".
Chromium has so much stuff packed into it, its insane. All that utility comes with Electron. And that's a good thing.
If you ever worked with video, for example, you know that having the full power of a modern browser in a desktop app is a game changer. Video playback (not to mention transcoding, which is also possible with modern web and webcodecs) is a complex beast, implementing that yourself is massive undertaking, not to mention in a desktop app that is supposed to work on win/mac/lin. I've built apps with Electron in tens of hours that would otherwise take me tens of days or more (and thats with AI because I'm not a video expert).
chromium is basically operating system at this point, it lacks kernel and ability to boot independently (added in chromium os), which is both good (from abilities standpoint) and bad (when copy of chromium is bundled with every app that renders webform with text field and button and nothing else)
when it goes about using webapps as desktop apps, native PWA support should be used, it would - in theory at least - lessen most issues electron apps have but will need extra effort and that's why we can't have nice things (like RAM free for other tasks)
> chromium is basically operating system at this point
I get what you're trying to say, but Chromium is far from being an OS. What you could say is, Chromium is as complex as an OS, or is replacing the OS in providing functional libraries atop devices (OS-provided application framework, if you will).
You are correct. Notwithstanding, people have been expressing the gp's sentiment for like a decade now [1] as is evident in this sub-thread [2], so it's a losing battle trying to prevent people from making such comparisons.
Those issues are typically from the decoding libraries you choose. Which could even be ffmpeg if you wish. GStreamer just provides a nice high level API.
> Web tech is just web tech. Yes it will allow you to render a button, but even unstyled, the button won't necessarily look native to the OS, and will vary between browsers.
Unfortunately nowadays even the built-in apps on the major desktop OSes are inconsistent, so the temptation for third-party apps not to care is somewhat understandable.
I change clothes all the time too, still match the pieces of clothing each time.
There's aesthetic value to coherence. There's also design, usability value. I have Telegram, Steam and Firefox opened right now and each one of them displays different minimize/maximize/close buttons on the top right. That's not ideal.
> They consistently miss the mark in adopting UI patterns from their host OS.
What you suggest is a disadvantage is one of the key advantages of Electron to me. I precisely do not want my things to look different on different OS. I don't have the resources to test my apps on all devices, and knowing that whatever I test on one system looks the same on another is A+.
The alternative is not that users get a UI specific for their OS, it is that users get nothing since I do not have the resources to develop for multiple systems. So yes, it is also an advantage to users.
Since when did anyone ever complain that youtube, google maps, roblox, or any other web sites didn't have native buttons and UI patterns?
Are you implying that the Windows, Mac, and Linux native desktop user interfaces don't all totally suck??! Or that there wasn't a huge celebration when Alan Dye finally left Apple for Meta? Or that users are clamoring for Jony Ive's infamous shallow superficial visual elegance over affordance and discoverability and usability?
Is it just too confusing for people to use youtube because the buttons don't look and feel exactly like native Mac buttons on the Mac and native Windows buttons on the Windows and whatever the kids are using on Linux desktops these days, therefore nobody uses youtube, and that it will only ever get popular if it just had a native look and feel?
Basically you're saying websites are the same as apps, and whether they're used in the browser or as a desktop app, the UI is fine to ditch that of the OS. Fine if you think so. I'll be sad to see OS and apps diverge completely in terms of UI.
YouTube succeeds for its content. Its UI is hot garbage both in the web and their apps. Google Maps is an atrocity and I’m very thankful Apple has decent data where I live. Roblox I don’t know, other websites I consume mostly in Reader mode.
I genuinely wonder who ever want that, and what apps those people use on their PC. Can you imagine, for example, Blender Foundation says that their next goal is to make Blender's UI look more like the host OS?
I have used Blender just a bit, but it was very jarring when I first opened it and discovered that it has its own menu bar within the window, rather than at the top of the screen. It has its own save/open window rather than using the system-provided one, as nearly all truly native Mac apps do. I doubt most Mac users like this.
I have somewhat more experience with QGIS. It has a standard Mac menu bar, but the icons are inelegant and Windows-ish, window layouts are Windows-ish, dialogs don't behave correctly in Mac full screen mode. It could use a MacOS glow-up.
I think Visual Studio Code (native menu bar, native save/open, but a core UI kind of unto itself that is consistent across Mac and Windows) does a better job of balancing cross-platform vs native.
And then there's the approach taken by Adobe and Microsoft Office. These apps do a much better job of adopting native platform appearance and conventions (sometimes at the expense of application consistency across platforms).
So... it has a menu bar where it belongs: on the application's main frame. But yes, that'll be jarring for Mac users, who have suffered a single menu with split-personality disorder for far too long.
As for the common dialogs, I agree with you. There's no better example of why you should stick with common dialogs than the shitshow that is Microsoft Office. The file-save... thing is mind-bogglingly bad. It looks like a hastily-thrown-together debug screen. You have no idea what you're looking at. The fields are primitive, unnecessarily-huge box outlines. There's no treeview to show you where you're working in the filesystem. There's a big list of what, history? Why?
At this point the only OS with a consistent look and feel at all is Mac. For the other OSes, I don't even know what a "native" look and feel would be. And most apps have their own branding and style they want to tout anyway. So I don't think "apps should look native" is the leading reason to not use Electron.
For me, the leading reason to use Electron is the fact that I already have a browser running so why not just use that to render your webpage... Make it a PWA if you want it in its own window.
Seems like I'm part of a shrinking minority (in this thread at least) who believes that web[sites/apps] in a browser, and apps running on the host platform, are different things in terms of UX expectations.
> At this point the only OS with a consistent look and feel at all is Mac.
Apparently, NOT! Apple's controversial Liquid Glass UI disrupted their long-standing Human Interface Guidelines by prioritizing cinematic aesthetics (transparency, refractions, and blur) over core tenets like readability, contrast, and content deference. Critics highlight that transparent, light-bending panes frequently clash with underlying wallpapers or videos, causing text and buttons to become illegible. This is a HOT topic in recent WWDCs and amongst UI designers.[1]
Tauri is a good way of packaging offline PWAs. It kind of sucks for building proper applications. The entire model for integrating with a local backend is just bonkers (no, I don't want my local application to pretend to be a web server - it's not a web server, I want to give you actual host bindings and share memory).
This. It's nuts how the whole industry accepts that typesetting engine from 80s with bunch of hacks on top is currently dominating cross-platform UI development.
I understand historical usage of the term with regards to desktop, but im unsure how you could consider HTML, CSS, JS, and runtime as anything other than a UI toolkit. They control the layout, styling, and actions of a UI.
- We can embed an existing application using a sidecar [1].
- Now, we can also use Elixir in the backend, embed the BEAM, and deliver a single binary, see ElixirKit [2].
As far as I know, LiveBook Desktop [3] is using Tauri for building binaries for MacOS and Windows. If Tauri works for the Elixir team, I think it works for me too.
Also, I know that Tauri is not bullet proof. WebView can be limited for some use cases, see [4]. There is some effort to use CEF to mitigate those problems, though [5].
I'd like to know how Deno Desktop compares with Tauri in this context. I know it is a new product, not sure if we could bundle an existing binary in Deno Desktop, like in ElixirKit.
Isn't all uses native OS UI widget? But since the brand need to be experienced the same across platform, it overrides the native rendering and use custom styles instead.
In practice it's much harder to maintain a native app. I am noticing this with ChatGPT Mac app vs. Codex Mac app. ChatGPT on Mac is constantly behind compared the web ChatGPT while Codex is shipping features at a much higher velocity.
Also ChatGPT hangs and has more weird bugs compared to Codex.
The issues with the ChatGPT Mac app could also be reflective of the state of Swift UI considering not even Apple themselves can ship Swift UI apps that aren't janky.
This ship has long, long sailed. If you don't spend your all your 24 hours as an office worker using Microsoft software, or you're locked in with a PC from 30 years ago, chances are almost every single UI you use will look differently, besides some microscopic agreements, like back button or a burger menu.
We just got used to it. There is some very vague thin layer of "commonly accepted patterns and symbols", but otherwise users just get through it.
We were writing and shipping desktop applications with it back in the nineties. Although many of the arguments against it were similar to the arguments against Electron today.
I think the UI look and feel was very ugly for many users and that caused its demise. The cross platform skin was ugly. The native skins were in the uncanny valley.
The framework was reasonably good for its time. By the time good looking UI frameworks came, the bad reputation was already set.
Even the later JavaFX was a tasteless exercise. I opened some apps and you could tell within 1 second that something was wrong because all the text was using fugly non-platform-native (or somehow screwed up) text rendering.
Compose and AOT compiled binaries would be amazing (GraalVM Native Image kinda thing) but it doesn't look very easy at the moment. Leyden with a regular JVM might be the best we get.
Compose UI apps can be compiled to native binaries already, via Kotlin's LLVM backend, though at the moment only the macOS/iOS targets have proper (official) support. Last I looked (a few years ago now), the Linux and Windows targets shouldn't be too far off, since it's all built on top of Skia already, someone just needs to care enough to put in the work. (But since right now you already get coverage for all platforms between JVM and Wasm, not to mention hot reload support on the JVM, there's little motivation to do so.)
I will die on the hill that Java was a good language, and had the potential to leapfrog us from where we are by at least a decade.
But it got hobbled by the awful, awful enterprise style culture, cultural misunderstanding of OOP (especially inheritance), and corporation shenanigans (fucking oracle).
I think this is a selection bias speaking rather than a reasonable reflection of what goes on in the Java world. Some insanely sophisticated and high quality technologies are written with Java.
The problem is like with JS or PHP, it is ubiquitous in many settings. There are a lot of people who can use it because it was the default language taught in CS programs, many corporate settings for decades, or similar. It’s the runtime for android devices. It’s everywhere. Of course you’ll encounter a lot of low quality developers.
Your comment mostly indicates that you haven’t been fortunate enough to encounter the high quality Java devs, not that they don’t exist. They exist and they build world class software that backs massive systems like elastic search, Kafka, spark, or Cassandra.
When I started to use elastic search I found out that with some incorrect queries you could "poison" the process entirely and it would respond incorrectly to every single query from that moment on, until you killed it and restarted it.
They responded to my issue several years later. I had changed jobs and I couldn't care less any longer.
It's all so bleeding edge right now. It also depends how deep you want to go. An increasing number of languages support wasm as a compile target, which is helpful.
Bytecode Alliance do semi-regular streams on Youtube. I think reading (recent) material on WASI (0.3) and the Component Model would be a good start.
Understanding the relationship between a host and a guest is valuable. Learning what wasmtime is and how it works is also illuminating: https://docs.wasmtime.dev
I think the last time I tried Deno for desktop it didn't allow for fullscreen webview apps. that was a showstopper for our kiosk apps. I'll have to revisit that issue and see if it's resolved now. I'm glad to see Deno continuing to march on.
You may have played with a 3rd party/unofficial solution or similar in the past as Deno Desktop has only just now become available in the Canary branch.
Having deno desktop do the framework handling for a bunch of popular options is an interesting choice. It seems deno is trying less to be an agnostic JS runtime, and more an "integrate everything toolkit" (not unlike Spring in the Java space).
I wonder if it supports opening invisible browser windows and doing things like intercepting cookies. In my desktop application I leverage a hidden browser window to manage auth state and use it like a proxy for the rest of the application. Might try to port it to deno desktop.
> The default WebView backend uses the operating system's own webview for small binaries, and you still have the entire npm ecosystem available through Deno's Node compat layer.
Sounds like a similar architecture to Tauri, but your business logic is in typescript instead of rust.
This works well. I’ve been using it to bundle other binaries with my applications and so far my users have had no issues on Windows, Linux, and macOS. I’m still a bit surprised given how new it is
This is something that can bring me back to Deno. Binary size is still pretty big IMO. If they can trim down runtime based on the specific uses of standard library (sort of treeshake Deno runtime itself), it would be revolutionary. Java promised this with `jlink` but sadly failed to deliver for wider adoption.
Of all the content they put out I liked the comparison section the most. The last row says iOS/Android - Electron: no, deno: not yet. If they deliver on this it will get much bigger.
This is kind of exciting. I have a lot of web development experience but every time I've tried to write a desktop app in the past, it just feels like a very clunky and unintuitive experience.
Smart move from the Deno team to get me to try out their ecosystem. I probably wouldn't have bothered prior. I've been mostly fine with npm, as its been much faster of late, and the security features recently released are good.
Curious to know who is using Deno in anger most days and in production full time? It seems like the choice of JS runtimes exploded over the past few years with that, Bun, etc.
In case someone is using something we haven't heard of e.g some are running using cloudflare workers which also has some unique runtime properties. AWS has something called LLRT.
I use it for several applications (frontend, backend, CLIs) in production settings, and it has been excellent. Caveat: I serve small internal teams mostly, some projects only serve < dozen users. One is around 500/day. No issues at all. I’ll definitely use Deno desktop for these internal tools. Their binary compilation (especially now that it can include other binaries) has been totally sound and I expect this to work well too.
Worth noting is that the team has improved compilation features steadily. Every issue I watched last year has been completed and I’m not encountering blockers anymore.
Impressive work. This is going to be really interesting for vibe coding Desktop apps. I imagine this on Lovable, Bolt or v0 since they basically default to using Typescript for building web apps. I've been using Go/Wails for desktop projects rather than a bundled Chromium and Node in a small desktop app, Electron did a good job but that was a big No for me.
Interesting but IMHO as we see on mobile providing WebViews work. Maybe instead of having Electron, Tauri, Electrobun, now Deno desktop but also plenty of alternatives then desktop browsers should provide WebViews on desktop with sandbox and permissions that make those applications usable. The alternatives listed here would just be fallback for a transition period until the WebViews are "good enough".
That's a whole can of worms, Micro$lop entangling its own browser with its OS, getting a (gentle) slap on the hand for its abuse of monopoly position for it, having to remove it claiming it's "impossible", etc.
Any indication when v2.9.0 will be released? I'd love to play with this, but the install instructions (at least for linux) don't seem to work. Is it working in Windowsland or on macOS?
IIRC Electron hello world is ~ 100-150 MB because it bundles a browser/Chromium runtime.
So I hoped we could have a <= 20 MB solution by reusing the OS webview or similar. Having more than 400 MB is a bit deceptive for me. (Again: maybe I just did something wrong in the config: should I do something else than `deno desktop test.ts`?)
libcef is the Chromium embedded framework[0], so your build isn't using a webview or maybe its using both. I just tried it on my mac, and I can't keep libcef out even with `--backend webview`.
I've decided on using a Clojure/Flutter hybrid that gets the best of all worlds. May integrate move from Bun to using Deno here https://codeberg.org/arik/clutter
Howdy, the Clojure dialects such as ClojureDart as well as Jank interop with the Flutter to provide a better development experience as well as more capable engine. This is inspired by Toyota's Fluorite implementation within flutter, which may be integrated into Clutter on its release.
I was just today morning thinking about some such idea, for mhtml to be embedded with light weight renderer so it does'nt have to rely on other browser
Maybe it is because it is still in development, but building and running the Hello World example just gives me a blank terminal and a white window that is not responding.
No matter how much you hate everything about js desktop apps — there are no proper alternatives.
The web is probably the closest thing the software industry has to a truly universal, open application platform. There is corporate influence, but it is substantially more vendor-neutral than any other UI platforms.
The web stuff mostly uses licenses such as MIT, Apache 2.0, and BSD. GPL-licensed projects exist, but still many more on permissive side.
Web is based on open standards developed through organizations and specifications are publicly available, royalty-free, and implemented by multiple independent browser engines rather than being owned by a single corporation.
can i open a socket with these tools ? can i open an odbc connection for example ? Or have i need to have a backend ? On desktop usually you can do much more than in the sandbox of a browser . I ask because i don't know these technologies. On windows, even if i don't like it, if the interface and the logic are not too complex, powershell with winform make you create things without "anything" installed and you can easily interact with other windows programs ( office 365 suite, autocad and so on ) so for doing "fast things" in my opinion is a very strong alternative .
I'm very inexperienced with regards to "JS on the desktop" environments. Does this mean any kind of improvements for Electron apps? Is it possible to port them to Deno?
While I've never liked to use deno compared to node and bun, this looks particularly good. The zero config options are nice, all the features seem to be in the place I like, and I'm happy they're not dogmatic about using the system webview and let you ship your own CEF. The state of system webviews on non-windows platforms is horrendous.
How does this differ from electrobun, which they explicitly mention, but make no point about? I had a quick drive with deno desktop and don't see how it's better. If anything it's lacking in comparison in my opinion. But hey, we can build desktop apps with deno now, too. So they got that going I guess ...
The world is trying to make computers faster and more accessible, more web UI slop isn't going to help that. Dumping Javascript entirely is the first step on that road.
I've seen variants of this comment for many years. The alternative to "web UI slop" would presumably be one of the many native toolkits.
I see it in a different way. The fact that "web UI slop" has managed to make great inroads on the desktop is an indictment of the state of native toolkits. If you think it's a problem that desktop apps are being written with web toolkits, the solution for that isn't to shame (as the term "web UI slop" clearly tries to do), but rather to figure out how to improve the native toolkits.
The opportunity to improve those toolkits was always there, and the ball was dropped.
It hasn't made any inroads on the desktop though... all anyone did was just package their own SWA into a self-contained browser that serves its own content. They continue to be websites, with all the pitfalls of them.
I don't need to spend 2GB-4GB of RAM just to have a over-glorified IRC clone!
Also, the native toolkits are fine. Windows has two toolkits, the ShellUI/MFC family (which does everything required, although it doesn't always get hidpi on legacy apps correct; it gets integration for blind people and also unicode/multilingual correct, and also works with touch interfaces), and WinUI does it more modernly (and ticks all the boxes). OSX has its toolkit, seems to nail everything correctly. Linux has Qt (lets ignore GTK for now, only reason you use GTK is if you want to appear Gnome-native), and Qt also does native++ uplifting on other toolkits (ie, native widget + additional feature expansion, plus perfect mimicry of native look for entirely new widgets), plus Qt does everything you need to do correctly and easily.
There are also new UI toolkits coming up through the ranks that are trying to knock Qt off that #1 position. None of the WebUIs would even place in this race.
Web UI toolkits always look non-native, are hard to interpret, often use low contrast (and frankly ugly) colorschemes, are easy to use in ways that do not comply with usability standards across OSes, and usually do nothing for A11Y.
The opportunity to improve those toolkits was always there, and the ball was dropped.
I think the fact that you listed off five toolkits for three different OSes, all of which are "that OS's own toolkit," might point at the root of the problem here.
Windows is so fucking old that I think it has a right to try again.
And, btw, the reason Microsoft even bothered is because (dun dun dunnnn) lots of internal apps at Microsoft were written with Qt, not MFC, and leadership got pissed when they realized (they couldn't tell the difference, since Qt does the native++ technique).
As for Linux, yeah, thats a shitshow. Qt was closed source, Linux isn't, so they made Gtk, but then Qt got open sourced, and then the Gtk guys just kept going. Qt can mimic Gtk3/4 themes and already uses 100% native rendering (same APIs Gtk does).
People keep writing Gtk apps for some reason, and I don't know why. I can get the C++ hate, but Qt has the cleanest C++ I've ever seen.
When I'm on Linux, I use KDE exclusively now (although I'm hoping Cosmic replaces it someday). The apps that use QML are fine even on decade old CPUs with absolute shit iGPUs, still better than even absolutely awfully designed WinUI apps.
Yes, native UI toolkits are not perfect, even though I consider Qt very close to one (I'm sure naysayers will find nitpicks). In the end, the choice is between the apps that eat 1GB of your RAM and learning to deal with some idiosyncrasies of native toolkits.
There's also just shipping with a Web interface that opens in a browser (like Jupyter, or WebUI apps). Plus there's the option to use the system Webview like Deno Desktop (this), Tauri and Electrobun do by default.
So thankfully we can still have our REPLs with live reloading and nice documentation (MDN, W3schools etc.) and large library of embeddable UI components without most the costs of using electron.
Even if SwiftUI, Qt and whatever Windows uses this morning were absolutely perfect, developers would write web UI slop to not have to write three frontends. That, and familiarity with JS are the whole reason.
You should try the last few Xcode versions some time. As far as I'm aware there's not a single line of Javascript in it, and all UI is 'native' (whatever that means these days), and the whole experience is such a janky and laggy mess that even VSCode feels slick.
It's trivial to write slow UI application in any tech stack, and just being 'native' really means nothing nowadays.
I challenge you to replicate something "simple", like an iMessage-style UI/UX in pure Win32 or WinForms.
The web views are entirely about productivity gains, not technological excellence. I don't know of many who would argue that the web view approach is superior from a purely technical perspective. It is mostly downsides in terms of performance, E2E latency, startup delay, security edges, etc.
Doesn’t mean you should use it for new things in 2026, but for building a chat app with? Yeah, it’d be very easy.
I’d know, because I’ve done it (granted it was over 15 years ago at this point, but I’ve done it). Was used internally, with a PHP backend. Worked great!
I watched traditional GUIs for a while and used many of them, mostly via ruby as wrapper, sometimes also via java. I finally reached the conclusion not too long ago, that web-apps are the only real alternatives now. Too many things do not work when it comes to traditional desktop applications. There are even regressions, e. g. ruby-gtk4 barely works for me. And there is no real support for any problems really. People make fun of e. g. electron "soooo big so bloated", and WebAssembly is still not really in any breakthrough after almost 20 years. But traditional desktop apps are also even more dead. So I'll have to add JavaScript/TypeScript/Node now simply because there are no real alternatives to this anymore. I'd wish we would have a real "write once, run everywhere"...
deno desktop ships in Deno v2.9.0 and is not in a stable release yet. To try it now, run deno upgrade canary to install the canary build. The command, configuration keys, and TypeScript APIs may still change before the feature is stable.
Vastly easier to set up, optionally lets you use platform web renderers, Typescript by default, you can use the Deno API instead of Node (compatible with less code but much better designed), built-in auto update, you can use Fresh which IMO is the best web framework.
RE: Tauri not having cross-compile... There's a GitHub action that compiles for Linux, Windows, and Mac. So practically it does have it, just not out of the box.
Practically that's just the ability to generate binaries for more than one target. "Cross compiling" is specifically that ability without having to invoke a separate external environment to get the additional targets.
If cross compiling were really just about the result rather than the means, what would the difference be between that and normal support for multiple targets?
what happens when two apps need different cef versions? doesn't that just mean you're back to bundling your own browser anyway. does the shared runtime actually save memory when the underlying chrome versions diverge?
> Shared CEF runtime across apps. Every app currently bundles its own CEF copy. A managed shared runtime would drop binary sizes to a few MB per app. On the roadmap.
This[0] sounds interesting. I am not familiar with CEF, so I wonder how the versioning works. When different apps require different versions of CEF, do we just essentially end up with the electron model where every app bundles their own browser (just slightly less bad). Or is there still an advantage to a "shared runtime" in that case?
[0]: https://docs.deno.com/runtime/desktop/comparison/
In case anybody else wondered CEF is the Chromium embedded framework.
https://github.com/chromiumembedded/cef
The biggest weakness of a framework like Tauri is the choice to target system webviews instead of bundling a browser runtime.
It seems great to be able to cut hundreds of megabytes out of your app installer, but the platform differences wind up being a complete and ongoing pain in the ass.
Tauri support on Windows is phenomenal.
Tauri on Mac runs into lots of WebKit/Safari issues, especially on older Mac machines that have an older engine that doesn't support modern web APIs. Your app can crash or be left non-functional. You'll find out about these runtime bugs in the wild randomly, and patching for some customers can take days, if not weeks.
Linux support is hellish, and it's best to not even try targeting Linux with Tauri.
Tauri is in the process of adding CEF support. It should probably become the default build target for all platforms.
I use Wails which is Tauri but for Go and I don't have the kind of issues you're mentioning. Maybe that is a difference between Wails and Tauri but I don't think the system WebView is a significant factor.
Are any of your Mac users using an 10-year old WebView? We frequently ran into that. And there's nothing that can be done about it except engineering around it.
I also doubt it works well on Linux. The performance of webkitgtk is like running an emulator inside an emulator.
Can you either at build time or runtime specify a minimum macOS version? No one running macOS 26 (for example) is using a 10 year old WebView.
But a lot of people are still running macOS < 10.15
Regular Tauri app (aptakube) user on linux here: the experience is very adequate and smooth, I have no complaints. Speed benefits relative to Electron (similar app: K8S Lens) alone are enough to deal with many possible issues.
Could be attributed to app developers going the extra mile, but I suspect it's the framework choice.
Targeting and building Tauri apps for Wayland, specifically, is a massive headache due to assorted webkitgtk bundling/incompatibility madness.
Web developers already have to deal with different browsers, versions and API coverage.
The beauty of shipping a desktop app would be to not have to deal with that though… also those built in web views are usually older than average.
This point of view always confuses me, because web developers already need to deal with platform differences. Especially if your app app also runs in a browser, like Slack and Discord — at that point, what issues do the differing system webviews cause that you don't need to deal with anyway just targeting browsers?
It's also funny to me as someone who's been building websites for 20+ years at this point, because the platform differences used to be much, much worse. Coincidentally, I just saw this article, which makes the case nicely: https://www.bram.us/2026/06/21/do-websites-need-to-function-...
webkitgtk isn't just quirky, it's also much slower compared to more popular browser engines and is particularly bad with RAM usage.
is webkitgtk different than the engine used in safari?
Not necessarily? As far as I know they come from the same WebKit source. But the bridge between WebKit and the OS is of course specific to each system. WebRTC is still experimental in webkitgtk while caniuse says Safari supported it back in 2017.
I remember encountering one bug with Final Form that triggered rarely in Safari, 100% in Webkitgtk, and never in Chromium.
Here's the developer of Tauri saying it's hard to recommend webkitgtk / Linux support:
> So if you need good linux support now/soon i can't 100% recommend tauri (for Linux) as of now. (I used to be more "forgiving" but with webkitgtk getting worse/more unstable each release i changed my mind)
https://github.com/orgs/tauri-apps/discussions/8524#discussi...
> This point of view always confuses me, because web developers already need to deal with platform differences.
On Mac, I use Firefox and Chrome.
However, if I use a Tauri app on Mac, I have to use dated WebKit. It's out of a Tauri developer's control.
> On Mac, I use Firefox and Chrome.
Sure, but many people use Safari, which runs that exact same WebKit engine under the hood. So if your app is available in the browser in addition to Tauri, you have to support it anyway. And at the very least, you as a web developer should be used to supporting it.
Many people run older versions of Mac OS X (because their machine won't upgrade), so their safari is aging -- they use an alternative browser, firefox and chrome support older mac os x.
Of course using unsupported OSes isn't the best idea, but Apple give you security updates longer than they give you 'fun' safari updates.
Sure, and many people also run older versions of macOS and also use Safari.
The platform webviews are significantly older/worse than typical web browser versions, especially on macOS and Linux.
On macOS, the only way to upgrade your WebView is to upgrade your OS, which requires rebooting. Lots of people just don't bother. You can upgrade to the latest Chrome or Firefox just by downloading it (assuming they support your macOS version), and they auto-upgrade themselves pretty aggressively.
On the web, very old versions of Safari (6+ years old) are a tiny fraction of a percent of your traffic; many web developers just ignore them. In a desktop app based on a WebView, ancient WebViews can be as high as 10-15% of your macOS user base. Ignoring them is not an option.
On Linux, it's common for the major version of WebKitGTK to not upgrade at all except during major OS upgrades. Anyone on Ubuntu LTS 20 is going to have a 2020-vintage WebKitGTK with security patches. (And Ubuntu LTS 20's WebKitGTK was buggier than macOS WebKit, even in 2020, because Apple has more dedicated full-time developers and testers making sure that macOS WebViews work end-to-end.) If you're shipping an app based on WebKitGTK, you can expect to see double-digit percentages of your Linux users running really old WebKitGTK.
Maybe you're such a great developer that your web app works great on ancient browsers, but, if so, it's probably because you didn't need/use much JS in the first place. (Maybe you used HTMX or something.) In that case, is there even any benefit in shipping a "desktop app"? What's your desktop app even for? Offline support? (But your app is all server side…?)
If you have a JS-intensive app that works great on ancient, buggy browsers, then platform WebView might work for you. It's not nobody, but it's hardly anybody.
Yes, we're used to platform differences, but that's kind of the benefit of shipping a desktop app, isn't it? So that we don't have to deal with the platform differences. If you bundle the browser, then that ought to all go away. No graceful degradation necessary, no polyfills.
LOL , then back to electron. I raised that since day one of Tauri.
Tauri's cef-rs is indeed quite good. I think it's available, ready, works: it's just that there aren't many folks using Tauri and only a fraction of them are aware/interested in exploring further. https://github.com/tauri-apps/cef-rs
Huge fan, this should definitely be the default. The user experience is incomparable.
One thing I'd like to verify: can the OS effectively use shared mem for cef across multiple different cef-rs apps? I really hope so. In this time of RAM being scarce, this optimization could be such a benefit.
Or use QtWebEngine which uses Chromium under the hood, provides a stable and rich API and enables sharing rather then bundling the browser runtime + is Cross-Platform + security updates are handled too.
Web devs are used to their target being evergreen, so I suppose you could opt in or out of that model: "just give me what you got".
> Web devs are used to their target being evergreen
I would think/hope web developers are used to “just give me what you got”. Any other mindset leads to “you must install <browser> to see this site”.
It’s Electron devs that are used to that.
I love how we’re now reinventing the browser as a much worse version of itself. What if instead of one or two general Web browsers we make everyone install 10 random versions that can only open one website?
> Web devs are used to their target being evergreen
Since when? The browser landscape is much better today than 10 years ago, but no web dev worth their salt assumes anything about the user agent.
True for this decade, but in the previous decade it was very much the opposite. Before you used any kind of browser api or nice language feature you would feature-detect it:
And if it wasn’t there you would define it yourself, it was called “polyfilling”. This was so commonplace that we built significant tooling like babel to standardize feature detection tests and fallback implementations - for a few years you could write
And behind the scenes the Rube Goldberg machine would turn this into something that would run in a JavaScript environment that had neither arrow functions nor promises.
System WebView but Electron is now the system
Ah, that's a new systemd module now.
I doubt the benefit. Practically every Electron app on a desktop uses different versions of Chromium and many are very out of date because of the risk of breaking when upgrading.
People build web apps for an array of browsers and huge ranges of versions. I think if you started using some tech to deploy an end user program and knew from the beginning the browser could be updated beneath you it would work just fine. But if you start with a golden version of Chrome and put off updating for too long you’ve let yourself get too comfortable.
> People build web apps for an array of browsers and huge ranges of versions.
en masse they don't. They just target the latest Chrome
De facto they do because functionality built three years ago and tested then is running along side functionality they built yesterday and tested on today’s Chrome.
People also do seem to test on iOS Safari because that pain in my ass needs special care on my software. So if a site works on it they either got lucky or tested on an iPhone. It’s generally only other people’s weird tech demo stuff that doesn’t work.
I agree and disagree, you can't target everything, but most (not shit) devs will target at least Safari - 1 or 2, simply because the iPhone market is too good to miss out on. And Safari being, well, Safari, means targeting that is a pretty safe bet for anything else.
Depends on the region, no one where I work has an iPhone or a current Mac, so stuff gets tested on FF and Chrome, and Safari gets thoughts and prayers. We would test on Safari if it were simple, but alas.
Skipping testing on 15% of all devices to save $600? Sounds like a poor business decision.
15% of devices is not 15% of users. From my own experience having a web app that is 99% desktop windows users, why would I care about safari?
Maybe for your app, it doesn’t make sense. And if it’s a pure enterprise app, fair enough (assuming it’s an enterprise that was started more than 15 years ago and only targets regulated or very specific markets). But a good way to guarantee that your app will never go beyond Windows desktop users is to ignore the most dominant mobile platform by users who actually pay for software.
Well, if your app doesn't run on Safari, you're not going to have any users on Safari.
We are a very small company, and have always had far more Firefox than Safari users. And though they get by via dominance, IE style bundling of the browser to the OS is toxic, so good riddance.
I target IE6 and it just works everywhere
You could, but by targeting a specific Electron app the mindset would be much simpler. Just take a look of how many times does the dev behind VS Code decide to upgrade their Electron/Node.js version, and how many breakages due to them.
It is all about unknown unknowns.
Just to let you know, CEF was used for Riot and League of Legends client as well [0]. The results haven't been nice, but I'm not aware if this was a problem with the CEF technology itself or other component/processes are to be blamed.
[0]: https://www.riotgames.com/en/news/architecture-league-client...
When the new client was built, microservices were the hot new buzzword.
The new client is some weird plugins/services based architecture. Things that'd barely warrant their own class in a boring OOP-based UI framework are instead now "isolated" services. The reality of this isolation being that if one piece breaks, the whole UI becomes unusable anyways. Dozens of things that in another app would've been just a simple synchronous call now behave like remote procedure calls and messages, forcing all the complexity of distributed systems into a local application for no reason.
That's why it runs like ass, breaks if you look at it wrong, and your CPU draws more power when using the client than when playing the game at 200FPS.
This is more a function of how mismanaged the project was at Riot (the iron client days, choosing Ember, etc.) which led to the current state of the launcher.
CEF is + has been the de-facto standard when you have a native app and want to do a web UI.
It's not the only option, but it's the most mature with the largest amount of docs + stack overflow questions, so it's a "safe" choice.
If you peek into the native resources files of most games/desktop apps, you'll find a good portion of them bundle + use the CEF dll.
I regularly install and uninstall the league client on Mac based on how I play that game (require fresh installs to raise the playing cost) and their client really sucks. Even freshly installing it, the client inhibits and hijacks mouse clicks from the full screen game when it's open. It took me months to figure out I would have to minimize the game, open and minimize the client (separate app) open, and then clicks would sometimes properly return.
Before that I was closing both and playing a game of roulette. By the the time my game was working, I may have already been reported for being afk and the game ended. Edit: other bugs have included starting to download the game and then signing in afterwards (their UI doesn't stop you) and then download progress disappearing, perhaps hanging, and you having no way to know it for a 40gb file unless over an hour has passed and you check disk size and then realize the client doesn't know how to load it. Start over and do a fresh install again, clear cache etc because their cache of the client still thinks somethings being downloaded even though it's not. Also having chat permanently off, results in weird glitches with friends requests and being unable to add new friends.
Horrible experience. But since the game is so optimized for addiction and dark patterns these days, and sunk cost, its a game I find myself returning to every once in a while.
On the gaming side, last I checked Steam's client was using CEF too and it doesn't get widespread blame for anything.
No shortage of games using it for in-game browser stuff, too.
Both Steam and Battle.net use CEF for their UI as well. And IMO they are on 2 ends of the "nice to use" from the implementation side (Steam being a sluggish hell and B.net being nice). Though then again B.net is only for Blizzard games, so they can also optimise for the limited set games.
> Steam being a sluggish hell
news to me. Been using steam since it launched. Never noticed it being sluggish
You haven't? Steam has been a miserable experience for years. Great way to make a monster gaming rig replicate the eMachines 1998 experience.
I'm so sincerely glad that 28 years later I get to spot hate for eMachine in the wild. What an absolute piece of shit. I remember buying one and immediately returning it and to this day... you get it.
Also Spotify. I believe Spotify was one of the earliest adopters of CEF, maybe the first major adopter?
The desktop app was originally a native C++ app, but they switched to CEF around 2011-2012. (It caused a very noticeable drop in performance!)
I'd prefer if it just used the system webview rather than downloading and managing an embedded browser itself. Webview2 on Windows for example.
> Small by default, full Node compatibility. The default WebView backend uses the operating system's own webview for small binaries
That appears to be the default, CEF is available if required (hint: it shouldn't be).
I used CEF for a project and Google is detecting CEF via some opaque algorithms and not allowing logins from it. From https://security.googleblog.com/2019/04/better-protection-ag...:
> Because we can’t differentiate between a legitimate sign in and a MITM attack on these platforms, we will be blocking sign-ins from embedded browser frameworks starting in June
Granted this was years ago, maybe the situation improved? I had to abandon my CEF project because of this.
Most apps (on desktop or mobile) open third party auth flows inside the user's default browser, which makes this a non-issue. For one, if you embed the Google login flow into your app then I can't reuse my existing session in my browser. But it also exposes my full credentials to your app for no reason, which is a good thing to avoid.
A strip down version of electron TBH.
I was wondering how this integrates with Deno's permission system, which is one of its biggest strengths especially for letting agents run amok on your device.
The CLI reference page[0] notes,
> The permissions you grant at compile time are baked into the compiled binary:
I think it would be nice if this could be surfaced to the user somehow, like letting the user know and decide which permissions they want to give access to.
[0]: https://docs.deno.com/runtime/reference/cli/desktop/#runtime...
> What deno desktop doesn't have yet
> Runtime permissions for desktop apps (a permission prompt on every filesystem / network access, i.e. Deno's permission system applied to desktop sandboxing).
You are running a binary that you got from the developer. If it presented you with Deno permissions, I think that would be misleading because there’s no guarantee of their integrity.
That is true. I wonder if it could be possible to let the user supply and wrap the app around their own, trusted installation of Deno (rather than the one bundled in the app) to specify permissions.
Why not run it in a vm or container instead then, it seems a bit much imho.
Then do `deno ./my_downloaded_deno_gui` instead. If you trust the copy of deno you downloaded, then hopefully it can be trusted to verify the permissions of random downloaded deno-apps.
Yes, it kind of defeats the standalone binary aspect but if you're really concerned about security, maybe it's a happy medium.
Deno continues to impress me. It’s honestly been quite a while since I started a new project without it. It has fully won my support over Node.js, the ecosystem has really matured nicely. I don’t know how often I’ll use this feature, but it’s really nice to have the option!
yes the big win from this is the Node API / NAPI support , if you write node_modules in nodejs , electron , raycast , edgejs you can reuse it .
https://wasmer.io/posts/edgejs-safe-nodejs-using-wasm-sandbo...
"NAPI allows native Node dependencies to target Node without actually depending on a specific version of Node or V8. NAPI abstracted the V8 JS engine away, by providing JS-like APIs for: creating an object, declaring a property, etc.
NAPI is the contract that all modern native Node modules use to interact with Javascript." [1]
1: https://wasmer.io/posts/edgejs-safe-nodejs-using-wasm-sandbo...
I'm still using node/npm and it's... fine. Every so often I read these posts and think I should change but node/npm is already a low friction part of my workflow.
However, what I have seen is that a lot of other libraries I use have switched to Bun. I haven't seen any that switched to Deno, and so I've been under the impression that Bun is becoming a strong node replacement candidate while Deno is not, or at least that the community is making a strong preference for Bun. Anyone have more insight into this?
This is a smart thing to ship. For me it would totally be a consideration when deciding on a platform to use.
agree, small footprint & cross-platform looks like a nice alternative to electron or tauri..
Their comparison page shows some savings, but not in every case (~40 MB / ~150 MB) https://docs.deno.com/runtime/desktop/comparison/
To be fair the ~150MB is for the CEF scenario (when chromium gets bundled); and eventually they want to introduce a shared CEF runtime across apps.
I'm happy to see this I see that this provides CEF, Webview and Raw * backbends but it would be nice if there was also a launch in browser option (like WebUI has). To me that has the best tradeoffs if you want to avoid the mess that is webkitgtk but still not ship (and be in charge of updating) a chromium engine with your app.
* https://docs.deno.com/runtime/desktop/backends/
> I'm happy to see this I see that this provides CEF, Webview and Raw
They beat Tauri at their CEF support.
Webviews are a mistake in most cases. They're too platform-specific, and certain Webviews (Safari/Webkit) are buggy as hell, making platform support a nightmare. (Linux, ironically, is even worse due to how underbaked webviews are on the major desktop Linuces - Tauri is barely functional on Linux.)
Deno Desktop could be a real contender in this space. It's good to see more Electron alternatives.
I think that my Sciter is better option when you need HTML/CSS/JS native application running on Windows (XP and beyond), MacOS and Linuxes.
Sciter SDK [1] contains scapp[.exe] - standalone Sciter engine that can be attached to HTML/CSS/JS bundle making standalone (single exe file) and portable executable. https://quark.sciter.com/ tool allows to compile such apps.
Size of "hello world" is a size of scapp.exe binary + size of compressed HTML/CSS/JS bundle.
On Windows scapp.exe is of ~14 Mb. On Linux ~18 Mb.
Linux version at startup detects GTK4, Wayland or X11 and uses those as windowing backends.
On all platforms Sciter provides out of the box: HTML/CSS/JS runtime, libuv based Node.JS alike runtime, GPU accelerated rendering, WebGL 3D runtime, JS built-in persistence (NoSQL DB).
It does not have TS compiler built-in as Deno, but that TS-to-JS compiler is better to be outside anyway as it is used only once - at app loading.
[1] https://gitlab.com/sciter-engine/sciter-js-sdk/
Wouldn't that just be "Raw"? I.e. start a webserver and ask the system to open the URL. There is no "special stuff" to do in this case like avoid sockets in favor of IPC to a well known webview or package CEF and no real integration to make with dev tools etc after - it's just open socket and serve from prebuilt binary.
No. As I understand it, the Raw backed just gives you a Window with input handling and you have to embed something like Skia, WebGPU for the graphics. So basically you have make your widget library yourself.
Now you can just start a server with deno pretty easily and serve a website. But WebUI will actually also manage opening the browser window for you as well a make the communication between backend and frontend just like using a Webview or electron.
Raw still exposes access to Deno APIs & ability to call native code but it doesn't seem to assume it should give you a window out of the box or anything based on my testing (AFAICT you have to orchestrate most all of that yourself, Deno just won't get in the way with packaging other stuff like CEF automatically). If this is just some issue on my machine with the canary build though, the existing "Deno compile" should accomplish the same goal for a binary version without any GUI components at all.
Outside making it so you don't have to call to open the link yourself, I'm still not sure what could be integrated in the scenario. Deno can integrate with WebView because the WebView APIs were designed to allow external applications to have full control of the session. CEF (the Electron-like) approach works because Deno packages CEF & the APIs as part of the app itself, having even more control of everything. Browsers are meant to be in control of themselves or the user, and have had a long history of fighting malware trying to act otherwise.
> Bindings are not IPC. The Deno runtime and the rendering backend run as threads / processes inside the same address space (CEF) or coordinated process group (WebView). Calls go through in-process channels, and the backend dispatches them from its run loop. -- https://docs.deno.com/runtime/desktop/bindings/
I don't understand how the coordinated process group works. Doesn't that mean in this multi-process mode it must be IPC? Maybe the claim "shared memory space" is more an architectural description than an OS-level claim?
this is what I wonder as well.
My guess is that it's not using IPC on OS level, like D-Bus on Linux, but rather a supervisior process starts and orchestrates child processes as needed. And all these processes use a shared memory model.
Here's the CEF docs on processes: https://chromiumembedded.github.io/cef/general_usage.html#pr...
EDIT: ...and the CEF docs on IPC: https://chromiumembedded.github.io/cef/general_usage.html#in...
from what i understand after a quick look at the source is it uses a C ABI to communicate between the WebView/CEF "host" application and the deno runtime which is loaded by the host as a shared library.
marshalling of values back and forth between the JS/C++/Rust layers still has to happen but these are just straight C api calls in process under the hood so much less overhead than having to do serdes across a socket/pipe.
- https://github.com/denoland/deno/blob/main/cli/rt_desktop/li...
- https://github.com/littledivy/laufey/blob/main/webview/src/m...
also notable that deno has a very low overhead bindings layer for doing JS->C/C++/Rust/Native interop using v8 fastapi calls where possible.
I'm happy for competition in this space, specially because Deno can run true TypeScript directly and not just strip types like the current Node implementation.
With that said, this is going to eat a lot of Tauri market. Why would I use Tauri now? The 150mb of additional bundle size is just an extra 1 to 10 seconds of download time in most internet connections and you get a reliable rendering engine.
> and you get a reliable rendering engine
How is it more reliable than Tauri - aren't they both using the system webview?
Deno Desktop can bundle CEF (Chromium Embedded Framework) according to https://docs.deno.com/runtime/desktop/comparison
Deno desktop can use system web view OR embed CEF. Tauri is just system web view.
The benefit of Deno Desktop is it's like Tauri except for when you want it to be Electron???
This is a feature many apps actually need.
E.g. Tauri uses WebKitGTK on Linux, which has historically been slow, unstable, and frequently lagging behind the main WebKit project. This is enough of an issue that even Tauri is working on the ability to use CEF instead of the system web view in Tauri apps.
Things are generally fine on recent versions of Windows and macOS. The system web views on these platforms will be evergreen versions of WebKit or Blink. But if you want to support very old versions of Windows or macOS, you might choose to use CEF instead of wrestling with Safari-from-five-years-ago.
> Why would I use Tauri now?
You’re “backend” isn’t JavaScript.
Deno also just strips the type annotations when running TS code - at least by default. To get type checking you'll need to run via `deno run --check`, or use the separate `deno check` subcommand. No big deal since type checking and linting usually happens automatically in the IDE during development.
Good to know. Does it also preclude features like enums?
Bartek from the Deno here. Nope, we do support enums OOTB.
Huh, I was going to mention Node's `--experimental-transform-types`, but that was completely removed in v26: https://github.com/nodejs/node/pull/61803
Wow, this is sad to hear. So if I understood correctly, node is walking back from trying to run TypeScript? Pretty sad if so.
I hope it's a temporary step back to leap forward, in the future.
Only for non-erasable Typescript! The support for running erasable TS is here to stay. It's not ideal, but on the other hand this is paving the way for the TC39 proposal - if that ever progresses...
Tauri doesn't lock you in to one JS ecosystem. In fact, it doesn't require you to use javascript at all.
Also, we've had several developer framework startups get acquired -- Astro, Nuxt, UV, Bun, Vite. It doesn't exactly inspire confidence in a software that you want to last and give support for years.
I don’t understand this logic. The acquisition doesn’t guarantee the project dies and staying unacquired doesn’t guarantee the project continues.
The state of OSS funding is precarious. The acquisitions at least guarantee some runway for the maintainers. Maybe the acquirer has alternative intentions than to simply bankroll the project’s the way it was, but at least someone is paying to keep the maintainers fed.
After having worked in the JavaScript ecosystem for over a decade, I don’t think _any_ project of significant size is guaranteed to last or have support. We need to be thinking about resilience (_when_ the framework stops being supported), not pretending like using OSS projects without paying them should magically give you some support contract/assurances.
But that would be the same argument for using electron? Why use this and not electron?
the option to use system webview , electron don't provided that
If you want desktop and mobile builds.
Tauri 2.0 added support for iOS and Android builds as targets.
> Deno can run true TypeScript directly and not just strip types
What exactly do you mean by that? Because no js engine carries the ts types into the runtime as far as I know. Deno and nodejs both use v8 as the runtime. v8's internal types are not connected at all to the ts types regardless of the wrapper. The only difference might be when/if static type checking is performed.
I think they mean deno handles transpiling for you so there’s no visible machinery for this aspect of the program. It’s just convenient.
I assumed they meant more than that because nodejs does the same since v22.18 and prior would do it with a flag. But they mentioned type stripping, which is really all transpiling of ts does for most code.
Temporarily at a place with 10-15 mbps. 150 MB is around 1 minute.
I grew up on 30 mbps. >= 100 is all I need nowadays.
But 10 mbps and websites and downloads really start to take a while.
The more bits and bytes you save, the more people will be pleased with your stuff! Even if they don’t know what bits and bytes are, and just go based on impatience
>I'm happy for competition in this space, specially because Deno can run true TypeScript directly and not just strip types like the current Node implementation.
this is misleading. there is no "running true ts". you will always be running pure js (until someone actually develops a "true" ts engine), and deno does "type stripping" just the same. the only difference is that it bundles the tools and makes it transparent and config-free which is more convenient (although more rigid).
There’s no running TS. Everyone executes JS.
Also TS itself is going towards stripped types. There’s this proposal which might land on browsers some day:
https://tc39.es/proposal-type-annotations/
And in preparation you can use this setting on your tsconfig:
https://www.typescriptlang.org/tsconfig/#erasableSyntaxOnly
> Why would I use Tauri now?
The same set of reasons anyone made desktop apps before Deno desktop launched. Deno desktop is not the reason desktop apps are made. If someone really wanted a desktop app made before, deno desktop isn't going to suddenly make new ones start appearing. It will just be a new choice among the many.
The overall feature seems really solid, but I'm impressed they couldn't reduce the average package size further from 40MB even when not using CEF. I guess that wasn't a huge focus when developing this feature? Tauri and Dioxus can easily hit less than 5MB for package sizes.
I find the feature matrix comparison to be extremely well done and the sections beneath explaining advantages and disadvantages to be some of the best docs I've read recently.
https://docs.deno.com/runtime/desktop/comparison/
Deno Desktop is bundling the V8 JavaScript runtime so it can have JavaScript on the backend. Tauri uses rust for the backend and your browser's JavaScript engine for the frontend.
> Web technology is the most widely-known UI toolkit in the world.
Poor choice of words there IMHO.
The reason Electron apps get a lot of flak is because they are everything _but_ a UI toolkit. They consistently miss the mark in adopting UI patterns from their host OS.
Web tech is just web tech. Yes it will allow you to render a button, but even unstyled, the button won't necessarily look native to the OS, and will vary between browsers.
None of that changes whether it's a UI toolkit, which it surely is.
> look native to the OS
Is that a problem? A button with a legible label is a button. The host OS doesn't have to look exactly like the applications it runs.
Consistency is a large factor in any good design, UI design more so.
They have internal consistency. The iOS version looks like the macos version which looks like the web version, etc.
This upsets HN users but the rest of the world decided that apps looking like windows built ins doesn't matter.
It's more like developers decided - nobody asked the users.
ironically the only group of users I've found that actually care about native UI, are other developers and Mac purist.
I have seen users having trouble with pixel soup UIs. They may not think "This should be in a native toolkit", but they do think "How the hell do I subscribe to a folder in the new Outlook?".
Right, but bad UI's was not uncommon before webviews, if anything the spartan-ness of the web often simpified patterns whilst reliance on weird hotkeys in desktop apps isn't uncommon.
>reliance on weird hotkeys in desktop apps isn't uncommon
The only examples I can think of are actions that are intentionally not easy to reach. How exactly it's done is platform-specific, but the (mis)feature doesn't come from the platform and can be implemented in other ways on other platforms.
- Apple UIs hide some power user functionality behind obscure hotkeys
- Similar: Shift-Delete to permanently delete (Windows, KDE) - Similar: Shift-right click to "Open With..." (Windows, KDE)
- In desktop apps that FOR SOME REASON try to look more like web apps, the hidden menu bar can be restored with Alt or Alt-M (Firefox, KDE)
The problem in these usability cases is pretty much always layout and constant redesigns rather than the exact theme the button has. I've seen plenty of unusable native ui soup UIs and very clean and simple custom UIs.
You could call it the exact theme when a clickable UI element looks just like regular text (it was not inside any kind of button-like shape in the Outlook case that I saw), but it's super common to have that in web-based UIs.
(And of course, nobody changed any theme in the outlook case)
> ironically the only group of users I've found that actually care about native UI, are other developers and Mac purist.
One group of people who routinely carry the can for poor product usability and another who by definition care about the Mac platform; entirely what would be expected.
That’s why HN users constantly advocate for Vim, a program in which every single thing works completely differently from every other modern application.
Yes, if there's one lesson from historical UI research that still holds, it's that mode switching is expensive. That's why people install vi plugins everywhere.
Wait...
Vi plugins don’t even exist for the vast majority of applications.
The last resort is to switch to ROU mode (key combo for entering it is :wq)
Consistent like what? Like maybe a decade ago one could say that osx was consistent, but nowadays even SwiftUI and cocoa is visibly different, let alone every second app that uses electron. And people don't care.
Windows has like 4 frameworks available on a bare new, latest OS install, just go deep enough in the "settings" or whatever they call it, and you can reach down to winforms. And on top the start menu is a react element!
(And in Linux you have the gtk and the qt world, and everything else)
> SwiftUI and cocoa is visibly different
Do they render different looking buttons?
Within OS consistency is much less of thing a thing than Web design conventions. Windows by itself has had several different UI frameworks over the years, so different "native" Windows programs can look completely different from each other.
> Within OS consistency is much less of thing a thing than Web design conventions.
Sorry, are you saying that two random web apps will typically share more UI consistency than two random Windows apps? Because, although I'm not currently familiar with Windows, I would be amazed if that were true.
Consistency is the reason why Electron is great. Consistency of the UI across operating systems.
If you think operating systems have nothing to offer in terms of UI patterns or guidance, then yes, that's a different type of consistency.
Great for the developer. The user doesn’t use Mac, Windows and Linux. Just one for work and one at home, with mostly different apps, so they couldn’t care less if it looks the same on different platforms.
They may care, however, if they get anything at all. I do not have the resources to target something to all platforms, so the alternative wouldn't be "Users get UI targeted towards their OS", the alternative would be "Users get nothing since developers don't have the time to also target their system".
> I do not have the resources to target something to all platforms
Some speculate that agentic engineering will enable the return of native apps
> I do not have the resources to target something to all platforms
What resources is actually needed? More often than not it just requires good engineering. You do not have to duplicate everything across platforms.
Time. I have the time to maintain a single GUI. I do not have the time to maintain three GUIs. Of course they'll be 99% the same, but checking that this 1% difference behaves and looks fine accross systems adds a substantial amount of effort that I simply can not support. And for what exacatly? I want them to be identical accross systems, not different.
> And for what exacatly? I want them to be identical accross systems, not different
For your users. Engineering is about designing things for the convenience of the builder, but for the convenience of the user.
A foolish consistency with terribly designed shallow superficial desktop user interfaces dreamed up by overpaid cocaine addled corporate boutique brand designers with not only no experience but actual burning contempt for usability and human factors and accessability and affordances is the hobgoblin of little minds.
https://daringfireball.net/2025/12/bad_dye_job
Yes, Dye botched [mac/i/tv/etc]OS 26.
That doesn't say anything about the value of whatever UI kit is in place, being shared consistently by apps. A virtue that, apparent from this thread, is no longer universally shared.
OS-level consistency is also consistency. It depends what we value. A lot of apps’ design could’ve been basic, OS-like UI. Apps such as GymBook or WhatsApp are internally consistent while still adopting many elements from the system’s design, instead of reinventing the wheel.
There are two types of consistency in this context: consistency within an OS and consistency across them. I, too, prefer the first because I only really use one OS, but this preference varies. I don't think it's right to say that the first case = "ui toolkit", but the second case doesn't.
Consistency as a design virtue applies to a single user's experience. This means consistency within the OS.
How is it a poor choice of words? It might not be "native" UI, but they never claimed as such.
I've always felt that native UI on Linux always looks incredibly ugly and I'd much rather use a nicely styled HTML+CSS layout instead.
In my experience, Electron mostly gets flak for being bloated and slow, it not being native is sometimes a secondary point people add on top.
I've always wanted to build a direct-browser integration that could use HTML+CSS for the layout, but avoids needing a JS runtime. Idk how lightweight servo is but one day I hope I will see my idea come to light
I don't mind the idea of using HTML components and widgets to style desktop applications. CSS and the DOM are widely known and reusing those for desktop apps is probably a good idea.
The problem, as you've pointed out, is that electron apps are bloated and slow. If they became the default and my editor, chat client, terminal, and everything else that I keep open were just thin layers around web applications, I'd rather figure out a way to move things into a browser rather than pull a piece of the browser out to wrap these applications.
There’s a point where it stops scaling in the browser, whether it’s due to scale or poor practices. For example Slack is annoyingly slow to start up and work in my FF, but works OK as an Electron app.
This is pretty much what Tauri solves. It re-uses the systems WebView. So the built apps are tiny (kb to a few mb)
> I've always wanted to build a direct-browser integration that could use HTML+CSS for the layout, but avoids needing a JS runtime. Idk how lightweight servo is but one day I hope I will see my idea come to light
Blitz (https://github.com/DioxusLabs/blitz) is exactly that. It's a new custom browser engine supporting standard HTML/CSS with a native Rust API (and optional integration with Dioxus which is a React-like UI framework in Rust). Baseline binary sizes are around 10mb.
We share a few components with Servo (Stylo the CSS engine (also shared with Firefox), and html5ever the HTML parser), but we've built a bunch ourselves too: notably we have our own layout engine, DOM tree and event handling. Servo is unfortunately tightly coupled with SpiderMonkey, and there is little prospect of removing that dependency in the short term.
That's really nice. It would be really good for game GUIs too where the situation is quite poor and would work well with underlays/overlays/worldspace UIs. That said while binary size may be around 10mb, it still baloons to 500mb at runtime for your TODO list example which is more than some electron apps.
Yeah, the RAM usage is quite high atm. The good news is that it's almost all the rendering layer (WGPU+Vello). So if we can make the renderer more efficient then it's likely we can bring that down. There is also some low hanging fruit in the DOM implementation (but I think that's only actually causing a few 10s of MB of usage).
I would also note that native toolkits (SwiftUI, etc) tend to also use at least ~100mb RAM these days. A lot of that is unavoidable if the app is actually visible, due to modern screen resolutions being so high.
This isn't exactly what you were suggesting, but it made me think of https://hacks.mozilla.org/2026/02/making-webassembly-a-first... since that article is about not needing to go through js to use wasm.
Every time I use Zed across Linux, macOS and Windows , I'm amazed stable and performant it's GPUI framework is. As a user, I'm very happy with it; of course some basic features like accessibility is missing but I'm sure it will be implemented soon.
As a developer, I'm not sure what's the barrier for entry is apart from Rust then again it's the USP as well.
Live reload. GUI development in compiled languages is a pain compared to web development.
Try dioxus, it has live reload but it's a work in progress.
Dioxus seems to be 'just' another way to generate HTML on the desktop. Electron but Rust? Is there a legitimate upside there?
With Dioxus, program logic compiles to native code instead of running it through a JS engine, and it ships its own HTML renderer (Blitz) instead of bundling a whole browser. So it's a lot more lightweight and performant than Electron.
As a minor bonus, the live-reload is also faster than what frameworks like React do. It truly has subsecond latency, which isn't exactly a game changer but is nice when iterating on visual details of an app.
Sounds similar to Wails. It does the same but with Go instead of Rust.
I don't think so, wails is more like tauri rather than diouxus
We now also have a custom GPU renderer (Blitz), which makes Dioxus more comparable to Flutter or the other Rust GUI toolkits (GPUI, Iced, etc).
For me it's not stable. After they change their renderer from one to another it freezes for me from time to time. But on the other hand I'm running Nvidia on Wayland so I feel no hate towards Zed owners - and restart is super quick ;)
Interesting, I'm on Nvidia in Wayland most of the time too and haven't had single freeze or crash in a very long time. Even the Windows is running inside the Qemu.
What DE? I'm on KDE.
> basic features like accessibility is missing but I'm sure it will be implemented soon.
Accessibility implementations frequently are more complex than the entire UI drawing bits. Most custom UI toolkits never implement accessibility, even decades after creation.
That hope is misplaced.
I agree about the complexity, but the core developers understand that the accessibility support is crucial and that's where my hope comes from.
[1] https://github.com/zed-industries/zed/discussions/6576
That's good news, but I would continue the timer there. That issue is from March 2023 so the counter is at 3 years, 3 months. Let's see if they have anything decent before 10 years.
Yeah, it is mostly laziness and cost cutting at the expense of users.
Nowadays there isn't even an excuse anymore, just vibe code it away in native frameworks.
When I can modify my desktop/theme (KDE) with css I will happily start doing it since that sounds easy.
Good news!
https://9to5linux.com/kdes-new-css-based-style-engine-union-...
Styling every application independently because it’s all individual Electron UI without a shared toolkit is much better indeed…
Which native framework?
Even in a "post-vibe code" era I wouldn't want to create multiple versions of the same app, and none of the "platform-native" GUI toolkits run on everything.
SwiftUI is apple-only, gtk has pretty bad compatibility on non-linux, qt is decent but requires C++ or python, and even so still not much for mobile. Don't even get started on "Windows frameworks", because as I write this sentence they may have left a new one in the ditch.
Flutter may be the closest, but why didn't they go with e.g. Java instead of a new language?
So yeah, if you want a truly universal UI then web is your best bet.
> if you want a truly universal UI
Right. If you want your app to look the same, custom way, ditching what the OS has to offer.
Some developers still believe an operating system has useful UI components and patterns worth adopting. From this thread it's clear that there's plenty who don't. Personally I view that as a regression.
Probably many Electron users also view that as a regression, but a tradeoff worth making.
Well, maybe Java's AWT has been correct all this time.
Of course there is value in having "OS-native" buttons, transitions, windows etc. And many parts of GUIs are basically standardized. The problem is all the parts that are not, and have to look the same everywhere.
The OS-component-oriented approach was defensible in the days of desktop only (though personally I think it was a mistake even then), but it's not sustainable now. People want your app on their PC to look and behave like your app on their phone (whichever combination of PC and phone they happen to have), and that's a lot more important to them than having it look and behave like other apps on their PC.
> People want your app on their PC to look and behave like your app on their phone (whichever combination of PC and phone they happen to have), and that's a lot more important to them than having it look and behave like other apps on their PC.
They don't. No one generates GBs of simulation data and create plots on their phones, and this is what "app" I am working on is doing.
How about something like this https://github.com/gen2brain/iup-go? Still not released, but I plan to clean all todos in the next few weeks.
The one which OS has to offer.
Web is bad everywhere outside of the browser.
I want to have both linux and mac users (but maybe also android, ios, windows).
You clearly see the issue.
> You clearly see the issue
I don’t. VLC is available everywhere, so your requirement is clearly not a problem. Jetbrains is available on all major desktop OS.
Well, getting a hardware-accelerated blank buffer onto a screen to render video content is hardly the epitome of graphical user interfaces. VLC has very few and basic UI elements.
Jetbrains is a better example, using Java with Swing which is not a common choice. As seen from my other comment, I do think this is a good direction, but it ain't any more native than Flutter or for what it's worth an Electron app, none of these are "what the OS provide". FYI Jetbrains has to do quite a few platform-specific tweaks to make them better citizens on each platform.
Portable applications is not a recent need. The only requirement is to have a standardize interface and an implementation for each of the platform. Where you put that interface is an engineering skill.
VLC went with QT (which has done all the hard work) for the UI, and their own libraries for the media playing part. Other software like Emacs and sublime has implemented their own UI libraries. Some just ship libraries and others build UI for them.
The issue with Electron is that it brings a whole jungle and a gorilla holding the single banana that the devs actually need. And the dev flung the whole thing at the users. It's like establishing an iron mine, a steel factory and then pollutes the whole region when building a quick stone bridge would do. Because the only thing you know are suspension bridges.
VLC is not the example you're looking for. Written in Qt for desktop and their own libVLC wrapper for mobile. Yes, in the case of VLC, parent comment is right: you clearly see the issue. And a media player is a relatively uncomplicated piece of software, UI wise.
When I started programming, one had to repeat in Assembly the same application for each computer brand.
We are not that bad nowadays, it is a skill issue.
There are plenty of ways to have portable applications with native UIs without shipping a browser.
Somehow we managed to do it for decades and without AI writing the code for us.
If you want to ship a browser, I already have one, thus standard Web, with a daemon if it really must be.
I agree on the standard web point, but you still failed to reply to what is a native UI for every OS.
Portable frameworks exist, but they are at most native to a single platform.
one missing from that list: Slint, which i work on. runs on Linux, Windows, macOS and embedded, with app logic in Rust, C++, Python or JS.
You can use JS but it doesn't ship a browser engine, it renders with its own lightweight toolkit.
I never thought I'd defend Electron, but I'd rather use the bloated web UI than a vibe coded Qt/GTK version I'm positive will not have seen any human QA.
But GIMP! /s
I have better things to do than spend my time adopting UI for various different systems. If Electron gives me the option to easily create a UI that looks the same everywhere, then I'll pick Electron over anything else any day.
I'm completely sure your software is an accessibility nightmare.
Looking native has long left the station as an objection about a UI.
Like 25 years ago. Nobody gives a damn since Microsoft stopped giving a damn.
I'm not part of the Apple world, but I thought they gave a damn?
Nah they stopped caring as well. Developing an application for macOS is hell nowadays. I hate the state of things but both Apple and Microsoft dropped the ball. Linux is even worse, so yeah I don't see a reality out of this unless we actually create something that surpasses the web in all measures.
Liquid Glass says no, they don't give a flying fuck any more.
Liquid Glass proves they do care (otherwise they wouldn’t have gone to the trouble), they’re just bad at it now.
Depends on the tool. We (mac people) tend to prefer native toolbars and settings menus, but I would say the days of relying on a "native" textarea or button are now behind us.
The other thing I find most Mac people appreciate is a shared understanding of hotkeys and if your app goes against the norm, one of the first feature requests will be to add configurable hotkey support.
Unfortunately, Apple has dropped the ball with their newest native apps in regards to UX and it will take years for them to go back and improve things. The new OS this fall is aiming to start that process, but it will still be a band-aid in some respect.
That is not why people use Electron. The goal is not and never was to just be a "UI toolkit" and "adopting UI patterns from their host OS".
Chromium has so much stuff packed into it, its insane. All that utility comes with Electron. And that's a good thing.
If you ever worked with video, for example, you know that having the full power of a modern browser in a desktop app is a game changer. Video playback (not to mention transcoding, which is also possible with modern web and webcodecs) is a complex beast, implementing that yourself is massive undertaking, not to mention in a desktop app that is supposed to work on win/mac/lin. I've built apps with Electron in tens of hours that would otherwise take me tens of days or more (and thats with AI because I'm not a video expert).
chromium is basically operating system at this point, it lacks kernel and ability to boot independently (added in chromium os), which is both good (from abilities standpoint) and bad (when copy of chromium is bundled with every app that renders webform with text field and button and nothing else)
when it goes about using webapps as desktop apps, native PWA support should be used, it would - in theory at least - lessen most issues electron apps have but will need extra effort and that's why we can't have nice things (like RAM free for other tasks)
> chromium is basically operating system at this point
I get what you're trying to say, but Chromium is far from being an OS. What you could say is, Chromium is as complex as an OS, or is replacing the OS in providing functional libraries atop devices (OS-provided application framework, if you will).
You are correct. Notwithstanding, people have been expressing the gp's sentiment for like a decade now [1] as is evident in this sub-thread [2], so it's a losing battle trying to prevent people from making such comparisons.
1: 24-core CPU and I can’t move my mouse https://news.ycombinator.com/item?id=14733829
2: https://news.ycombinator.com/item?id=14736498
> Just as a data point - Chrome has more code than the linux kernel -
> It's an operating system (pretending to be a browser).
Chrome OS is literally an operating system that's 90% through chromium
gstreamer is not that complicated
It has a really really crappy security record, though.
Those issues are typically from the decoding libraries you choose. Which could even be ffmpeg if you wish. GStreamer just provides a nice high level API.
Electron doesn't?
Last time I checked there's a small industry of gstreamer contractors, so it's not that simple.
> Web tech is just web tech. Yes it will allow you to render a button, but even unstyled, the button won't necessarily look native to the OS, and will vary between browsers.
The irritating, and unnecessary, pedantry.
Indeed. Even Qt isn't native, in the most purist sense.
Who cares if it looks native?
Native UIs change all the time too and not always for the better.
Change over time is something different from apps looking vastly different at any given time.
Unfortunately nowadays even the built-in apps on the major desktop OSes are inconsistent, so the temptation for third-party apps not to care is somewhat understandable.
I change clothes all the time too, still match the pieces of clothing each time.
There's aesthetic value to coherence. There's also design, usability value. I have Telegram, Steam and Firefox opened right now and each one of them displays different minimize/maximize/close buttons on the top right. That's not ideal.
So if you already wear clown shoes does that mean that you have to wear only clown costumes?
Someone who is in full clown costume looks less ridiculous than someone who wears clown shoes and is otherwise dressed "normally", yes.
> They consistently miss the mark in adopting UI patterns from their host OS.
What you suggest is a disadvantage is one of the key advantages of Electron to me. I precisely do not want my things to look different on different OS. I don't have the resources to test my apps on all devices, and knowing that whatever I test on one system looks the same on another is A+.
That's an advantage to you. Not necessarily to your users.
The alternative is not that users get a UI specific for their OS, it is that users get nothing since I do not have the resources to develop for multiple systems. So yes, it is also an advantage to users.
Users don't actually give a shit.
This is a techie complaint, and that's opting for a charitably nice description.
> Users don't actually give a shit.
Have you ever had a job as a tech support? If not, you don’t have anything to say.
People do complain about inconsistent UX. Especially when it does not behave like the platform it’s on.
Yeah, I've done that job countless times in my 20 year career.
Knock it off already.
Since when did anyone ever complain that youtube, google maps, roblox, or any other web sites didn't have native buttons and UI patterns?
Are you implying that the Windows, Mac, and Linux native desktop user interfaces don't all totally suck??! Or that there wasn't a huge celebration when Alan Dye finally left Apple for Meta? Or that users are clamoring for Jony Ive's infamous shallow superficial visual elegance over affordance and discoverability and usability?
Is it just too confusing for people to use youtube because the buttons don't look and feel exactly like native Mac buttons on the Mac and native Windows buttons on the Windows and whatever the kids are using on Linux desktops these days, therefore nobody uses youtube, and that it will only ever get popular if it just had a native look and feel?
Basically you're saying websites are the same as apps, and whether they're used in the browser or as a desktop app, the UI is fine to ditch that of the OS. Fine if you think so. I'll be sad to see OS and apps diverge completely in terms of UI.
YouTube succeeds for its content. Its UI is hot garbage both in the web and their apps. Google Maps is an atrocity and I’m very thankful Apple has decent data where I live. Roblox I don’t know, other websites I consume mostly in Reader mode.
> UI patterns from their host OS
I genuinely wonder who ever want that, and what apps those people use on their PC. Can you imagine, for example, Blender Foundation says that their next goal is to make Blender's UI look more like the host OS?
I have used Blender just a bit, but it was very jarring when I first opened it and discovered that it has its own menu bar within the window, rather than at the top of the screen. It has its own save/open window rather than using the system-provided one, as nearly all truly native Mac apps do. I doubt most Mac users like this.
I have somewhat more experience with QGIS. It has a standard Mac menu bar, but the icons are inelegant and Windows-ish, window layouts are Windows-ish, dialogs don't behave correctly in Mac full screen mode. It could use a MacOS glow-up.
I think Visual Studio Code (native menu bar, native save/open, but a core UI kind of unto itself that is consistent across Mac and Windows) does a better job of balancing cross-platform vs native.
And then there's the approach taken by Adobe and Microsoft Office. These apps do a much better job of adopting native platform appearance and conventions (sometimes at the expense of application consistency across platforms).
So... it has a menu bar where it belongs: on the application's main frame. But yes, that'll be jarring for Mac users, who have suffered a single menu with split-personality disorder for far too long.
As for the common dialogs, I agree with you. There's no better example of why you should stick with common dialogs than the shitshow that is Microsoft Office. The file-save... thing is mind-bogglingly bad. It looks like a hastily-thrown-together debug screen. You have no idea what you're looking at. The fields are primitive, unnecessarily-huge box outlines. There's no treeview to show you where you're working in the filesystem. There's a big list of what, history? Why?
Garbage.
At this point the only OS with a consistent look and feel at all is Mac. For the other OSes, I don't even know what a "native" look and feel would be. And most apps have their own branding and style they want to tout anyway. So I don't think "apps should look native" is the leading reason to not use Electron.
For me, the leading reason to use Electron is the fact that I already have a browser running so why not just use that to render your webpage... Make it a PWA if you want it in its own window.
> Make it a PWA if you want it in its own window.
Seems like I'm part of a shrinking minority (in this thread at least) who believes that web[sites/apps] in a browser, and apps running on the host platform, are different things in terms of UX expectations.
> At this point the only OS with a consistent look and feel at all is Mac.
Apparently, NOT! Apple's controversial Liquid Glass UI disrupted their long-standing Human Interface Guidelines by prioritizing cinematic aesthetics (transparency, refractions, and blur) over core tenets like readability, contrast, and content deference. Critics highlight that transparent, light-bending panes frequently clash with underlying wallpapers or videos, causing text and buttons to become illegible. This is a HOT topic in recent WWDCs and amongst UI designers.[1]
[1] https://blog.prototypr.io/why-apples-liquid-glass-design-is-...
Tauri is getting traction in the meantime.
A non-native UI has some issues, but also one clear advantage - it is easier to make a cross-system app with the same looks.
Tauri is a good way of packaging offline PWAs. It kind of sucks for building proper applications. The entire model for integrating with a local backend is just bonkers (no, I don't want my local application to pretend to be a web server - it's not a web server, I want to give you actual host bindings and share memory).
This. It's nuts how the whole industry accepts that typesetting engine from 80s with bunch of hacks on top is currently dominating cross-platform UI development.
Nobody cares about this anymore.
On Windows you have 20 different ways to write native apps that all look different.
On Linux you have Qt/GTK and god knows what else.
Only macOS is somewhat consistent although with Liquid Ass it's also getting worse.
I understand historical usage of the term with regards to desktop, but im unsure how you could consider HTML, CSS, JS, and runtime as anything other than a UI toolkit. They control the layout, styling, and actions of a UI.
For me, the best about Tauri is that:
- We can embed an existing application using a sidecar [1].
- Now, we can also use Elixir in the backend, embed the BEAM, and deliver a single binary, see ElixirKit [2].
As far as I know, LiveBook Desktop [3] is using Tauri for building binaries for MacOS and Windows. If Tauri works for the Elixir team, I think it works for me too.
Also, I know that Tauri is not bullet proof. WebView can be limited for some use cases, see [4]. There is some effort to use CEF to mitigate those problems, though [5].
I'd like to know how Deno Desktop compares with Tauri in this context. I know it is a new product, not sure if we could bundle an existing binary in Deno Desktop, like in ElixirKit.
[1] https://v2.tauri.app/develop/sidecar/
[2] https://elixirkit.hexdocs.pm/tauri.html
[3] https://github.com/livebook-dev/livebook/blob/main/rel/app/t...
[4] https://www.youtube.com/watch?v=vmslGvxObvM&t=621s
[5] https://github.com/orgs/tauri-apps/discussions/8524
There's a comparison page: https://docs.deno.com/runtime/desktop/comparison/
And you can embed anything on the binary with `deno desktop --include […]`.
As much as I like cross-platform stuff, I also really like native UIs that follow native UX patterns, etc.
We spend a lot of time using different browsers. As far as I know there no web engine that use native OS UI for rendering.
Isn't all uses native OS UI widget? But since the brand need to be experienced the same across platform, it overrides the native rendering and use custom styles instead.
> As far as I know there no web engine that use native OS UI for rendering.
That sounds like a monster I would be afraid to touch.
In practice it's much harder to maintain a native app. I am noticing this with ChatGPT Mac app vs. Codex Mac app. ChatGPT on Mac is constantly behind compared the web ChatGPT while Codex is shipping features at a much higher velocity.
Also ChatGPT hangs and has more weird bugs compared to Codex.
Did they run out of tokens? Why don't they ask their agent to update the mac version?
The issues with the ChatGPT Mac app could also be reflective of the state of Swift UI considering not even Apple themselves can ship Swift UI apps that aren't janky.
https://daringfireball.net/2026/06/swiftui_only_makes_it_eas...
This ship has long, long sailed. If you don't spend your all your 24 hours as an office worker using Microsoft software, or you're locked in with a PC from 30 years ago, chances are almost every single UI you use will look differently, besides some microscopic agreements, like back button or a burger menu.
We just got used to it. There is some very vague thin layer of "commonly accepted patterns and symbols", but otherwise users just get through it.
Why did they describe electrobun as macOS only? I checked their docs and it has support for Windows, macOS and Linux
https://docs.deno.com/runtime/desktop/comparison/ https://github.com/blackboardsh/electrobun#platform-support
Thanks, I'll update the docs. When we wrote them a couple weeks back, Electrobun was announcing Linux only support.
I swear we're just going to end up with Java again.
At this point I think that would be a more sane outcome than whatever it is we have right now.
We were writing and shipping desktop applications with it back in the nineties. Although many of the arguments against it were similar to the arguments against Electron today.
I think the UI look and feel was very ugly for many users and that caused its demise. The cross platform skin was ugly. The native skins were in the uncanny valley.
The framework was reasonably good for its time. By the time good looking UI frameworks came, the bad reputation was already set.
Even the later JavaFX was a tasteless exercise. I opened some apps and you could tell within 1 second that something was wrong because all the text was using fugly non-platform-native (or somehow screwed up) text rendering.
I think SWT was the best option if you wanted native controls.
Yeah back then a java application could take several minutes to open.
Java would be a killer platform if they shipped built-in, tauri-like, UI.
There's Compose Multiplatform if you are willing to switch to Kotlin. Only caveat is that it uses Canvas rendering on web.
Compose and AOT compiled binaries would be amazing (GraalVM Native Image kinda thing) but it doesn't look very easy at the moment. Leyden with a regular JVM might be the best we get.
Compose UI apps can be compiled to native binaries already, via Kotlin's LLVM backend, though at the moment only the macOS/iOS targets have proper (official) support. Last I looked (a few years ago now), the Linux and Windows targets shouldn't be too far off, since it's all built on top of Skia already, someone just needs to care enough to put in the work. (But since right now you already get coverage for all platforms between JVM and Wasm, not to mention hot reload support on the JVM, there's little motivation to do so.)
Compose Multiplatform is just a worse Flutter.
The Revenge of Javagator!
https://news.ycombinator.com/item?id=19837817
I will die on the hill that Java was a good language, and had the potential to leapfrog us from where we are by at least a decade.
But it got hobbled by the awful, awful enterprise style culture, cultural misunderstanding of OOP (especially inheritance), and corporation shenanigans (fucking oracle).
I have nothing against java. But for some reason in my experience all the developers using it are low quality, and gave it the reputation it has.
I think this is a selection bias speaking rather than a reasonable reflection of what goes on in the Java world. Some insanely sophisticated and high quality technologies are written with Java.
The problem is like with JS or PHP, it is ubiquitous in many settings. There are a lot of people who can use it because it was the default language taught in CS programs, many corporate settings for decades, or similar. It’s the runtime for android devices. It’s everywhere. Of course you’ll encounter a lot of low quality developers.
Your comment mostly indicates that you haven’t been fortunate enough to encounter the high quality Java devs, not that they don’t exist. They exist and they build world class software that backs massive systems like elastic search, Kafka, spark, or Cassandra.
When I started to use elastic search I found out that with some incorrect queries you could "poison" the process entirely and it would respond incorrectly to every single query from that moment on, until you killed it and restarted it.
They responded to my issue several years later. I had changed jobs and I couldn't care less any longer.
If that's your example of quality… well…
If a single outlying case in a broadly used piece of software is your example of an entire language's developer base, well...
A critical bug in your example of excellent software written in java. I've of course seen many horrors in other java projects.
Java may be good, but it's boring. No joy comes from programming in Java.
I need to enjoy my work to be engaged and productive.
You can use Clojure to get all the goodness from Java and still have fun
That is a matter of taste. I enjoy programming in Java just fine.
It is better to be D language then - natively compiled and has superb meta programming features.
Especially if to consider that I've added native D support to Sciter [1].
[1] https://terrainformatica.com/2026/06/05/ai-assisted-developm...
Similar to something I'm working on for games: https://jumpjet.dev
WASM you can bundle for Windows, macOS, Linux, Android, iOS and web. Unlike Deno Desktop, it doesn't rely on a browser engine.
Do you reccomend and resources for building w/ & learning about wasm?
It's all so bleeding edge right now. It also depends how deep you want to go. An increasing number of languages support wasm as a compile target, which is helpful.
Bytecode Alliance do semi-regular streams on Youtube. I think reading (recent) material on WASI (0.3) and the Component Model would be a good start.
Understanding the relationship between a host and a guest is valuable. Learning what wasmtime is and how it works is also illuminating: https://docs.wasmtime.dev
Tyty
Do you mean "Unlike Deno Desktop"? Deno Desktop definitely relies on a browser engine.
Yes, thanks.
Just FYI, when checking out jumpjets homepage, the white-dot airship in the background made the white text in the hero banner hard to read.
Cool project!
Deno Desktop supports two backends as of now: CEF (Chromium) and Webview
You can get your app sizes as low as 15mb with `deno desktop --compress` (in canary)
A tiny "raw" windowing backend exists for WebGPU rendering as well
I think the last time I tried Deno for desktop it didn't allow for fullscreen webview apps. that was a showstopper for our kiosk apps. I'll have to revisit that issue and see if it's resolved now. I'm glad to see Deno continuing to march on.
You may have played with a 3rd party/unofficial solution or similar in the past as Deno Desktop has only just now become available in the Canary branch.
Having deno desktop do the framework handling for a bunch of popular options is an interesting choice. It seems deno is trying less to be an agnostic JS runtime, and more an "integrate everything toolkit" (not unlike Spring in the Java space).
Looks actually good!
I wonder if it supports opening invisible browser windows and doing things like intercepting cookies. In my desktop application I leverage a hidden browser window to manage auth state and use it like a proxy for the rest of the application. Might try to port it to deno desktop.
> The default WebView backend uses the operating system's own webview for small binaries, and you still have the entire npm ecosystem available through Deno's Node compat layer.
Sounds like a similar architecture to Tauri, but your business logic is in typescript instead of rust.
> Opt into the bundled Chromium (CEF) backend when you need identical rendering across macOS, Windows, and Linux.
Sound more like Electrobun
Actually, this would be amazing for distributing web games as apps for Steam or online purchase. I am going to give it a try.
Neat! Is there any "bundle/integrate with existing native application" story like Tauris sidecar [0]?
[0]: https://v2.tauri.app/develop/sidecar/
From what I understood, what you want is `deno desktop --include […]`.
> Includes an additional module or file/directory in the compiled executable.
This works well. I’ve been using it to bundle other binaries with my applications and so far my users have had no issues on Windows, Linux, and macOS. I’m still a bit surprised given how new it is
That's great to know, thanks! Will look into that.
Is it going to support iOS/Android?
No customizable programmable browser runtimes exist for those.
What do you mean by that? Does this[1] not count?
[1] https://capacitorjs.com/
Correct, that isn't a browser it uses a webview. The limited webview version of Deno Desktop could work like that on phones.
Bartek from the Deno team here. No promises yet, but we're looking into feasibility of it.
The docs say it's planned but no proper roadmap for it yet.
This is something that can bring me back to Deno. Binary size is still pretty big IMO. If they can trim down runtime based on the specific uses of standard library (sort of treeshake Deno runtime itself), it would be revolutionary. Java promised this with `jlink` but sadly failed to deliver for wider adoption.
Of all the content they put out I liked the comparison section the most. The last row says iOS/Android - Electron: no, deno: not yet. If they deliver on this it will get much bigger.
This is kind of exciting. I have a lot of web development experience but every time I've tried to write a desktop app in the past, it just feels like a very clunky and unintuitive experience.
Smart move from the Deno team to get me to try out their ecosystem. I probably wouldn't have bothered prior. I've been mostly fine with npm, as its been much faster of late, and the security features recently released are good.
Curious to know who is using Deno in anger most days and in production full time? It seems like the choice of JS runtimes exploded over the past few years with that, Bun, etc.
Why "etc."? Isn't it just node, bun and deno? Genuine question
In case someone is using something we haven't heard of e.g some are running using cloudflare workers which also has some unique runtime properties. AWS has something called LLRT.
https://github.com/awslabs/llrt
https://developers.cloudflare.com/workers/runtime-apis/
We have created Edge.js that can run Node.js apps fully using your preferred JS runtime: V8 or QuickJS.
https://edgejs.org/
There's QuickJS, LLRT, Rhino, and GameMaker is about to get TS/JS support.
I use it for several applications (frontend, backend, CLIs) in production settings, and it has been excellent. Caveat: I serve small internal teams mostly, some projects only serve < dozen users. One is around 500/day. No issues at all. I’ll definitely use Deno desktop for these internal tools. Their binary compilation (especially now that it can include other binaries) has been totally sound and I expect this to work well too.
Worth noting is that the team has improved compilation features steadily. Every issue I watched last year has been completed and I’m not encountering blockers anymore.
Impressive work. This is going to be really interesting for vibe coding Desktop apps. I imagine this on Lovable, Bolt or v0 since they basically default to using Typescript for building web apps. I've been using Go/Wails for desktop projects rather than a bundled Chromium and Node in a small desktop app, Electron did a good job but that was a big No for me.
Interesting but IMHO as we see on mobile providing WebViews work. Maybe instead of having Electron, Tauri, Electrobun, now Deno desktop but also plenty of alternatives then desktop browsers should provide WebViews on desktop with sandbox and permissions that make those applications usable. The alternatives listed here would just be fallback for a transition period until the WebViews are "good enough".
Is it not a web view? With nodejs capabilities from the “backend” half of the app for normal desktop app filesystem access etc?
If a Web View is not provided by a browser then it's an already installed browser then it's as they say "web rendering engine" that they ship along.
I'm trying to argue that it should already be available via Firefox, Chromium, etc on desktop.
Firefox doesn't release a webview engine but if they did I wonder if Linux distros would use it.
Webviews have always worked since MSHTML, the issue is being comfy helping Google's market share instead of writing portable Web code.
That's a whole can of worms, Micro$lop entangling its own browser with its OS, getting a (gentle) slap on the hand for its abuse of monopoly position for it, having to remove it claiming it's "impossible", etc.
Any indication when v2.9.0 will be released? I'd love to play with this, but the install instructions (at least for linux) don't seem to work. Is it working in Windowsland or on macOS?
> deno desktop is opinionated about those tradeoffs:
> Small by default, full Node compatibility
I tried `deno desktop index.ts` with the 5-line Hello world in the article.
Result (Windows 10): 442 MB. Ouch.
I thought it would be smaller than an Electron build, but it's far worse. Did I do something wrong?
(libcef.dll: 247 MB) (deno-test.dll: 78 MB <- contains the hello world)
IIRC Electron hello world is ~ 100-150 MB because it bundles a browser/Chromium runtime.
So I hoped we could have a <= 20 MB solution by reusing the OS webview or similar. Having more than 400 MB is a bit deceptive for me. (Again: maybe I just did something wrong in the config: should I do something else than `deno desktop test.ts`?)
libcef is the Chromium embedded framework[0], so your build isn't using a webview or maybe its using both. I just tried it on my mac, and I can't keep libcef out even with `--backend webview`.
https://github.com/chromiumembedded/cef
Try the webview backend: `deno desktop --backend webview`
The docs say that's the default:
https://docs.deno.com/runtime/desktop/backends/#webview-(def...
apologies, this is inaccurate currently, will get things updated
I've decided on using a Clojure/Flutter hybrid that gets the best of all worlds. May integrate move from Bun to using Deno here https://codeberg.org/arik/clutter
Care to expand on what this solution is? What is Clojure bringing to Flutter in this case? Ty.
Howdy, the Clojure dialects such as ClojureDart as well as Jank interop with the Flutter to provide a better development experience as well as more capable engine. This is inspired by Toyota's Fluorite implementation within flutter, which may be integrated into Clutter on its release.
I was just today morning thinking about some such idea, for mhtml to be embedded with light weight renderer so it does'nt have to rely on other browser
> for mhtml to be embedded with light weight renderer so it does'nt have to rely on other browser
That's Sciter (https://sciter.com) and Sciter.Quark (https://quark.sciter.com) in particular, no?
> Backend and UI communication goes through in-process channels, not socket-based IPC
Are they running the frontend and backend in the same process? Sounds a bit dangerous security-wise?
How can in-process channels be more dangerous than a socket-based IPC? The frontend still goes through the "secure" JavaScript engine AFAICS.
Maybe it is because it is still in development, but building and running the Hello World example just gives me a blank terminal and a white window that is not responding.
Sorry to hear that, could I ask you to file an issue in our bug tracker?
So how much does a hello world weight?
Is this a new trend? Why are everyone starting to do desktop runtime? For example, I recently saw Bun Electron, and then I saw this project.
I guess people are tired of each instance of an Electron-based app using 1GB+ of RAM.
I just tried `deno desktop helloworld.ts` and the result is 442 MB. So it's not any lighter than Electron (see my toplevel comment)
Then it probably was wishful thinking on my side...
Why are people parroting this meme?
For 8 years now, constantly updating to newest Electron, my Electron app has been using only about 150mb of RAM - (see Video Hub App)
Hmm suppose you have a node GUI-less application. What would you pack it in to have something reasonably self contained to deploy?
Look into Single Executable Applications
https://nodejs.org/api/single-executable-applications.html
Thanks!
Deno actually has had a built in compile to binary feature * I've used it before a few times.
* https://docs.deno.com/runtime/reference/cli/compile/
No matter how good they get, I still hate everything about js desktop apps
I'm truly curious about running another desktop environment inside via WASM, honestly...
No matter how much you hate everything about js desktop apps — there are no proper alternatives.
The web is probably the closest thing the software industry has to a truly universal, open application platform. There is corporate influence, but it is substantially more vendor-neutral than any other UI platforms.
The web stuff mostly uses licenses such as MIT, Apache 2.0, and BSD. GPL-licensed projects exist, but still many more on permissive side.
Web is based on open standards developed through organizations and specifications are publicly available, royalty-free, and implemented by multiple independent browser engines rather than being owned by a single corporation.
Deno desktop supporting other backends using raw is crazy!
can i open a socket with these tools ? can i open an odbc connection for example ? Or have i need to have a backend ? On desktop usually you can do much more than in the sandbox of a browser . I ask because i don't know these technologies. On windows, even if i don't like it, if the interface and the logic are not too complex, powershell with winform make you create things without "anything" installed and you can easily interact with other windows programs ( office 365 suite, autocad and so on ) so for doing "fast things" in my opinion is a very strong alternative .
Yes, you can do all that. You get a fully-fledged Deno program that can do all of this, _plus_ you get a frontend GUI app.
Why this and not rust or go?
That would give us efficient desktop apps that don’t need a JavaScript runtime.
Literally the first question in TFA.
Awesome! Looking forward to trying this out.
I'm very inexperienced with regards to "JS on the desktop" environments. Does this mean any kind of improvements for Electron apps? Is it possible to port them to Deno?
They had this before I used it to ship some stuff but binaries were big . How small did they get it with this update
I hope bun desktop is coming soon?
I expect a poorly conceived and buggy vibe coded version will be available this afternoon.
While I've never liked to use deno compared to node and bun, this looks particularly good. The zero config options are nice, all the features seem to be in the place I like, and I'm happy they're not dogmatic about using the system webview and let you ship your own CEF. The state of system webviews on non-windows platforms is horrendous.
So I guess this is a competitor to Electron?
How does this differ from electrobun, which they explicitly mention, but make no point about? I had a quick drive with deno desktop and don't see how it's better. If anything it's lacking in comparison in my opinion. But hey, we can build desktop apps with deno now, too. So they got that going I guess ...
I think for a little while longer, you can catch bun anything, electro or not, refugeess just by not being bun.
Bartek from Deno team hear. I'll love to hear what you feel is missing so we could improve `deno desktop` for more users.
Funny how Deno desktop supports prompt(), which electron refuses to implement
I don't get the point of this.
The world is trying to make computers faster and more accessible, more web UI slop isn't going to help that. Dumping Javascript entirely is the first step on that road.
I've seen variants of this comment for many years. The alternative to "web UI slop" would presumably be one of the many native toolkits.
I see it in a different way. The fact that "web UI slop" has managed to make great inroads on the desktop is an indictment of the state of native toolkits. If you think it's a problem that desktop apps are being written with web toolkits, the solution for that isn't to shame (as the term "web UI slop" clearly tries to do), but rather to figure out how to improve the native toolkits.
The opportunity to improve those toolkits was always there, and the ball was dropped.
It hasn't made any inroads on the desktop though... all anyone did was just package their own SWA into a self-contained browser that serves its own content. They continue to be websites, with all the pitfalls of them.
I don't need to spend 2GB-4GB of RAM just to have a over-glorified IRC clone!
Also, the native toolkits are fine. Windows has two toolkits, the ShellUI/MFC family (which does everything required, although it doesn't always get hidpi on legacy apps correct; it gets integration for blind people and also unicode/multilingual correct, and also works with touch interfaces), and WinUI does it more modernly (and ticks all the boxes). OSX has its toolkit, seems to nail everything correctly. Linux has Qt (lets ignore GTK for now, only reason you use GTK is if you want to appear Gnome-native), and Qt also does native++ uplifting on other toolkits (ie, native widget + additional feature expansion, plus perfect mimicry of native look for entirely new widgets), plus Qt does everything you need to do correctly and easily.
There are also new UI toolkits coming up through the ranks that are trying to knock Qt off that #1 position. None of the WebUIs would even place in this race.
Web UI toolkits always look non-native, are hard to interpret, often use low contrast (and frankly ugly) colorschemes, are easy to use in ways that do not comply with usability standards across OSes, and usually do nothing for A11Y.
The opportunity to improve those toolkits was always there, and the ball was dropped.
I think the fact that you listed off five toolkits for three different OSes, all of which are "that OS's own toolkit," might point at the root of the problem here.
Windows is so fucking old that I think it has a right to try again.
And, btw, the reason Microsoft even bothered is because (dun dun dunnnn) lots of internal apps at Microsoft were written with Qt, not MFC, and leadership got pissed when they realized (they couldn't tell the difference, since Qt does the native++ technique).
As for Linux, yeah, thats a shitshow. Qt was closed source, Linux isn't, so they made Gtk, but then Qt got open sourced, and then the Gtk guys just kept going. Qt can mimic Gtk3/4 themes and already uses 100% native rendering (same APIs Gtk does).
People keep writing Gtk apps for some reason, and I don't know why. I can get the C++ hate, but Qt has the cleanest C++ I've ever seen.
Gtk is hardware (GPU) accelerated, while Qt Widgets is software/CPU-rendered.
I notice you say, specifically, Qt Widgets.
Yes, classic Qt Widgets still doesn't allow for hardware acceleration.
However, the majority of Qt UIs you deal with are in Qt Quick, which is hardware accelerated. Almost all of KDE is Qt Quick, for example.
Yes, I was considering C++ (and C) desktop UI toolkits. Unlike Qt Widgets, QML/JavaScript is a horrible mess and not suitable for desktop software.
When I'm on Linux, I use KDE exclusively now (although I'm hoping Cosmic replaces it someday). The apps that use QML are fine even on decade old CPUs with absolute shit iGPUs, still better than even absolutely awfully designed WinUI apps.
Yes, native UI toolkits are not perfect, even though I consider Qt very close to one (I'm sure naysayers will find nitpicks). In the end, the choice is between the apps that eat 1GB of your RAM and learning to deal with some idiosyncrasies of native toolkits.
There's also just shipping with a Web interface that opens in a browser (like Jupyter, or WebUI apps). Plus there's the option to use the system Webview like Deno Desktop (this), Tauri and Electrobun do by default.
So thankfully we can still have our REPLs with live reloading and nice documentation (MDN, W3schools etc.) and large library of embeddable UI components without most the costs of using electron.
Even if SwiftUI, Qt and whatever Windows uses this morning were absolutely perfect, developers would write web UI slop to not have to write three frontends. That, and familiarity with JS are the whole reason.
You should try the last few Xcode versions some time. As far as I'm aware there's not a single line of Javascript in it, and all UI is 'native' (whatever that means these days), and the whole experience is such a janky and laggy mess that even VSCode feels slick.
It's trivial to write slow UI application in any tech stack, and just being 'native' really means nothing nowadays.
I challenge you to replicate something "simple", like an iMessage-style UI/UX in pure Win32 or WinForms.
The web views are entirely about productivity gains, not technological excellence. I don't know of many who would argue that the web view approach is superior from a purely technical perspective. It is mostly downsides in terms of performance, E2E latency, startup delay, security edges, etc.
In WinForms?
Is that supposed to be a challenge?
WinForms is incredibly easy and simple to use...
Doesn’t mean you should use it for new things in 2026, but for building a chat app with? Yeah, it’d be very easy.
I’d know, because I’ve done it (granted it was over 15 years ago at this point, but I’ve done it). Was used internally, with a PHP backend. Worked great!
I watched traditional GUIs for a while and used many of them, mostly via ruby as wrapper, sometimes also via java. I finally reached the conclusion not too long ago, that web-apps are the only real alternatives now. Too many things do not work when it comes to traditional desktop applications. There are even regressions, e. g. ruby-gtk4 barely works for me. And there is no real support for any problems really. People make fun of e. g. electron "soooo big so bloated", and WebAssembly is still not really in any breakthrough after almost 20 years. But traditional desktop apps are also even more dead. So I'll have to add JavaScript/TypeScript/Node now simply because there are no real alternatives to this anymore. I'd wish we would have a real "write once, run everywhere"...
Odd because I'm wrapping up an app that uses Xaw. It should run on the billion or so machines that support X11
> I'd wish we would have a real "write once, run everywhere"...
Hi! I'm trying my best okay? Hokusai Pocket might currently be composed of wood and string, but one day he will be a real boy!
https://docs.deno.com/runtime/desktop/#hello%2C-desktop
Yeah, hello desktop.
D:\source\DenoTest>deno desktop main.ts
error: Module not found "file:///D:/source/DenoTest/desktop".
How is this better than Electron?
Vastly easier to set up, optionally lets you use platform web renderers, Typescript by default, you can use the Deno API instead of Node (compatible with less code but much better designed), built-in auto update, you can use Fresh which IMO is the best web framework.
These from the comparison table stick out to me:
- Can use "raw, system WebView, or bundled CEF" vs "bundled Chromium"
- The size can be smaller in the Raw/WebView cases.
- Built-in automatic differential updates
- Built-in cross-compilation (+the compilations just come built into Deno itself rather than as a 3rd party package).
And, of course, the same lists as one would generate when comparing the base Deno vs Node themselves.
would be cool to have a comparison with tauri.
They really did their best comparing it with other tools here https://docs.deno.com/runtime/desktop/comparison/
RE: Tauri not having cross-compile... There's a GitHub action that compiles for Linux, Windows, and Mac. So practically it does have it, just not out of the box.
Practically that's just the ability to generate binaries for more than one target. "Cross compiling" is specifically that ability without having to invoke a separate external environment to get the additional targets.
If cross compiling were really just about the result rather than the means, what would the difference be between that and normal support for multiple targets?
Enough with web trash! Isn't vibe coding enough to build a proper program?
Great, another way to make shipping bloated javascript apps easier. Just what we need.
Another Chrome wrapper...
what happens when two apps need different cef versions? doesn't that just mean you're back to bundling your own browser anyway. does the shared runtime actually save memory when the underlying chrome versions diverge?
This web page stole my scroll liberty