That sounds like an amazing idea. I've been using a lot of Rust+GPUI and sometimes behaviors that I'm expecting to be there just aren't.

Swapping it for UI that's as battle tested as Blender might improve things.

Super curious about the limitations you see in GPUI! I built a custom rust UI framework for the software i am developing but i always kept an eye on GPUI, what is limiting about it in your opinion?

I'm curious about the limitations too.

GPUI and Makepad are the two Rust UI frameworks that seem the most interesting for what I'm looking for in a UI framework (hybrid immediate/retained mode, self-drawn widgets, open source).

Not so much limitations. Rather behaviors I'm used to, around scrolling content, input fields, splitters for resizing panes etc. Can't recall specifics right now. Just small things that I usually get out of the box had to be implemented here.

A glaring limitation is that it requires a modern GPU, which makes it unusable in many contexts.

always liked blenders design language, looks cool!

Ive been thinking this would be amazing to do for ages. happy someone finally did it. its the perfect system for interfaces where speed matters. hope to see more 3d modellers, video editors, daws and simulation systems built on this.

thanks for sponsoring!! Do let me know if there's anything you built with it :)

It would be such a boon if GIMP and Inkscape could adopt blender's fantastic GUI

De gustibus et coloribus non est disputandum.

vel adrogantia

This feels like those AI models suddenly switching to Chinese mid-conversation

This entire thread caught me off guard lmao

Gustus est sicut nates - divisus est.

cum emptore

lorem ipsum

Lol ;)

expecto patronum

Inkscape’s UI is much, much better than Blender’s, IMO.

Inkscape is actually one of the few FOSS apps that is actually pretty intuitive to use. In fact, way back when I first started using vector design tools, I found it more intuitive/had better feedback than Illustrator (same thing can't be said about GIMP vs Photoshop). Inkscape really is great, but I have one huge gripe: I'm sure there's a system to it, but the file export name never stays what I want it to be...

Weird, I find it to be an unscrutable turd

With a much simpler set of tasks to accomplish.

People are confusing UX with UI. Inkscape UX might be better for many people but it would be hard to argue that Inkscape UI is better - it got better recently but it's still far from blender polish.

I'm surprised how much everyone's opinions differ in this thread. I've always found blender's UI incredible but inkscape's awful (though nowhere near as bad as gimp)

These complex tools are "intuitive" to whoever uses them the most

They are completely incomparable applications.

and then rename it blimp

Fantastic is the last word I would use the describe blenders UI/UX.

A side by side screenshot of inkscape/gimp/blender for those who don't have them installed

https://cleanshot.com/share/WW9yFdMy

One of the biggest problems with GIMP and Inkscape is how GTK renders numeric inputs. Look at the size of those +/- icons on the toolbar. For drawing software with lots of numeric inputs, that means an insane waste of space. Proprietary software like DaVinci and even Photopea have much smaller inputs, allowing them to create far more complex GUIs and featuring without cluttering the window.

Tbh, you could summarise a lot of the issues with their UI as "GTK"

GIMP is far from a nice example of what GTK can do…

For sure, which is ironic given it was the application GTK was created to support. But even so, I don't think I've ever really seen a GTK-based application I actually liked (the rest of the UI is going to have to make up for the atrocious file picking dialog, for example)

A lot of these UIs were built around icons, following the direction MS Office took in the early 2000s. These once had colours and were somewhat distinguishable.

When they tried to jump on newer trends they made them all monochrome, and they don't have distinctive sillouettes so it's just completely inscrutable. That Inkscape screenshot is horrible. I count 129 icons.

And yet they still devote so much space to their bizarre colour palette, which is huge but still manages to omit most of the gamut? It's the kind of design that really only comes from decades of OSS development.

I just want to be able to crop and draw on top of images. Like old Paint. That is literally all I want to ever do with images. What should I use in Linux for this purpose

GIMP of course.

KPaint

Use KolourPaint, it's a windows 9x/XP paintbrush clone with only a few improvements tacked on.

Photocraft has been working alright for me. I really struggled to learn GIMP, and I haven't found any easy lightweight image editing tool that replaces Paint.NET for me before trying Photocraft. It's still a little confusing since I'm not used to photoshop-like bindings and layout, but it hasn't been too hard to adjust. (I also only ever use it for small simple edits, drawing, cropping, color gradients, background removal, etc.)

There is paint.rs[0] as well which is intended to be paint.net open source alternative which works on Linux.

I’d be interested to know what you think about it

[0]: https://github.com/daijro/paint.rs

I wish people would stop naming software after the implementation language.

Thank you for showing me this!

I tried it out, and it's a pretty good clone. It serves my muscle memory well, with a few hiccups like drag-and-drop image onto the window doesn't work exactly the same, the save menu is a bit different. But there are also some pretty good improvements, like the addition of text layers (text got rasterized immediately after loosing focus in paint.NET which made it HORRIBLE to work with. This change is super welcome). There are also a bunch of new "plus" menu items that it seemingly gets from PhotoCraft? Not sure if that's because I have PhotoCraft installed, or if it comes with those features bundled. I'm not sure how good this is, because Paint.NET menus were already quite bloated and confusing to navigate, so this makes them even more cluttered with for example "Hue / Saturation" next to "Hue / Saturation +". The UI windows can't be dragged out of the paint.rs window like they could in Paint.NET. I think this is also faster than Paint.NET was.

I like this, it's quite good. I would like a more modern UI though, ideally using native GTK elements. Though that raises and interesting question of how much should a re-implementation veer away from the original. I think I'll keep this as a backup if I get fed up with PhotoCraft.

Perhaps Gradia [0], Pinta [1], or Krita [2].

[0]: https://gradia.alexandervanhee.be/ [1]: https://www.pinta-project.com/ [2]: https://krita.org/en/

Thanks for mentioning gradia. That seems perfect actually

I easily modified pinta to do exactly this. You mean overlay, right? It can overlay, draw, manipulate, and I added a better file navigator so I can see which imagery I want to overlay to begin with. My use case was scientific in nature. What is yours, if you don’t mind my asking?

IrfanView in Wine, f12 opens a really old-school paint toolbox.

https://photoflare.io/ is the one I've found to be most usable. Pinta is just unpleasant to use at high resolutions and it's buggy.

It's just pointless to port GIMP over at this stage, you may as well do a greenfield project without the cultural baggage and poor foundations.

I love the general idea behind Blender's UI.

And I do appreciate the effort that goes into it, especially in the latest versions as it got much better,

But I really really really REALLY don't want that UI in anything else other than Blender.

Please don't.

If you're working on a small game project without an existing framework/platform, you usually build a really bargain-basement game editor alongside it. That editor takes a long time, and it's often spent in arranging UI stuff.

If you're already using Blender to make assets, there's some benefit to stuffing Blender's UI into your game editor too.

After trying blender for about 20 years I still find the UI/UX awful. I think a lot of the new blender generation has baby duck syndrome.

> That's actually extremely common in scientific (GUI) software. And in any professional setting I guess.

I don't think that qualifies. There's an important distinction to be made between an aesthetic preference versus a professional not appreciating a tool being mucked with. Anything that interferes with my workflow without immediately providing substantial benefit is something that's actively harmful. In fact even if it does hypothetically provide huge benefits I still might not want to deal with it right this second.

......huh?

So if it's for an employer it's rational but if you're making the same choice for yourself it isn't? If one of those is baby duck syndrome they all are.

I think they're noting how the nature of baby duck syndrome might depend on whether the change introduced is within something that is for work or for play.

As in, that people who refuse change in their play activity are in some way different (less rational? more "syndromatic"?) than those who refuse change in their breadwinning activity.

I see the distinction, but am not sure I agree. It's just a neurotype spectrum of the same phenomenon. Baby duck is baby duck

Not quite. The distinction is between avoiding an articulable and concrete negative impact versus mere aesthetic considerations. For example, this pan is more difficult to use because it is somewhat undersized for what I'm trying to cook versus I don't like this pan because it's the wrong color or has the wrong accents or whatever. I understand baby duck syndrome to refer to the latter - ie having an unsubstantiated bias towards aspects of the thing first encountered.

I'd only point out that undersize and oversize aren't purely objective concerns, but are also aesthetics compared to what you're used to.

I wouldn't be arguing about this except that I can't trust my clock app on my phone because some bozo keeps redesigning it to work a different way every 6 months

And I think that's an excellent example. I don't think dislike of sudden and unexpected changes to your clock app would qualify as an instance of baby duck syndrome precisely because it has a material impact on your usage of it.

In contrast, if you prefer phones with blue bodies for no particular reason and your first phone happened to be blue then it's at least possible that's an example (or maybe you just really like that color, I dunno).

Blender had a very different ui before and a big redesign, so most people have different opinions between the old and the new design. It sounds like you find all of them awful. I find that surprising and wonder what 3d program you find has good ui.

I have used every version of blender and it has always felt harder than it should.

I have used 3Ds max, maya, lightwave, cinema4D, milkshape, wings, Houdini, unreal engine 1 - 5, Hammer, Threejs editor and probably more. The Blender UI is worse than all of the above.

The question was which one do you think has a good UI... I'd be keen to know your opinion on this as well, what you think the current high water mark is in 3D design software UI is, or do you think they all suck, but Blender just sucks the most?

At the time of blender's old ui 3Ds max had menus nested 5-6 level deep in an ui that would make gimp very good design. Between that 3Ds max praise, conflating old blender ui and new blender ui, and not providing any specifics on what makes blender so bad (I just provided some about 3D max) I have to assume there is no truth to it. We will agree to disagree. I think blender ui is excellent, but a 3d editor is a specialised app.

With 3Ds Max listed as good UI we really must just have different tastes, hah! :D

You've given quite a list of tools with different purpose.

For 3D work I would personally rate

Modo 3D, Zbrush > Blender, Houdini > Softimage XSI > Maya > space intentionally left empty :P > 3Ds max

Blender UI looks very nice and is very customizable. I think it's "ok" in beginner discoverability (none of 3D softs is great at that really) but when you get experienced with it it's one of the best.

I really like how they managed to keep all the tools and views pretty consistent given how many things you can do in it. Whether you are working on objects, vertices, nodes, movie clips you get the same shortcuts where possible.

Fascinating, I would have ranked Softimage XSI above blender (even modern blender), it was just that intuitive to me.

3Ds max being dead last is fitting, people who think it's peak UI just have stockholm syndrome in my opinion.

[delayed]

Apparently there is a crowd still using 2.79b, the last version with the old UI (current LTS is 5.2)

My baby duck fingers could probably pick up LightWave 3D again in an hour or two. I could never gel with Blender.

I'm not a professional designer or 3D modeler but I've used Blender quite a bit. I know what you mean but the latest UI post redesign isn't bad.

Blender does a LOT of stuff. Its not an easy product to create an UI for. But I think the problem is much bigger when coming from a different 3D package than when learning from scratch.

My teenage kids use Blender as the first 3D package they've ever learned. I was impressed with how quickly they ramped up through videos and learning from each other.

> I think a lot of the new blender generation has baby duck syndrome.

Maybe some of them do, but even though I disliked the UI at first, I've come to appreciate it as I've learned it. It's simply different from most other software, so that obviously causes friction. However, I think once you learn it, it's quite a good system. There's a tradeoff between optimizing for familiarity/new users, and providing an optimized experience for advanced users.

I agree, I don't think blender has changed as much as people say. But I don't think it is baby duck, I think blender always had a good UI, it has just taken the rest of the world 20 years to catch up.

A lot of people go "Blender UI was horrible, but they redesigned the whole thing and now it is good" Having used blender in a minor role for 20 years(i think I started on 1.7), I think this was just a big psyop by the blender team. Sort of how the client goes "Change things" and you change nothing, wait a day, and tell the client "I made the changes you requested" client: "I love it"

I am not saying the blender team has not made substantial improvements to the UI over the years, but the core is still that same quick blender workflow, This is very different from the normal clunky GUI design of 3d modelers of the era and as such blender got a reputation for having a weird UI which was easily translated to blender having a bad UI. So what do you do when you have an unwarranted bad reputation? Especially now that the rest of the world has caught up and is using your flows and they are not so weird anymore. Make a dark mode and put out a big press release "we redesigned everything, it's good now" so people will try it.

I feel like the 2.49 to 2.6 update was a bit overblown, but the 2.8 update was pretty damn massive to me eyes

In that case it would be competing with imgui, egui and similar libraries, but the submission is comparing it to GPUI.

[delayed]

>If you're working on a small game project without an existing framework/platform, you usually build a really bargain-basement game editor alongside it. That editor takes a long time, and it's often spent in arranging UI stuff.

Not in an age of AI.

I mean if you want to code all this stuff by hand, I get it. But that doesn't mean shoehorning a Blender UI into a "bargain-basement game editor" is a generally useful thing because it would take you a long time to arrange UI stuff without AI.

[delayed]

I think it’s just about as weird as questioning somebody’s motives on an Internet comment thread. I certainly share GP’s opinion about game tooling, but i’m here because i still care about making good software and that involves learning about and navigating technological developments as they happen. Is that sufficient reason for you?

[delayed]

You are reading them wrong. I think they mean “customer facing” instead of user facing i.e. the player doesnt care if the level editor has some jank: they never see it.

For the types of tools that your coworkers are paid to use, vibing a first draft and developing from there is so much more time and money efficient than writing by hand that you would be dumb to spend your tooling budget differently.

Thats how i read it. Someone still needs to use tools, no one is saying that claude is going to one shot mass effect 4 yet.

Depending on what you and the other poster mean by "arranging UI stuff", AI could either solve the problem for you or create larger problems, especially if you're not opinionated on either technical or UI matters.

Imagine this spinning off and becoming this generation's GTK.

Not going to happen, as it's GPL.

Akin to Java in Oracle vs Google case, keeping the API but LLM-ing the implementation can be considered as fair use.

There can never be a compatible LGPL release?

Only by reimplementing it all, as Blender is GPL.

So, about a week's worth of vibe coding.

The GPL made me pause.

There's this thing called Qt...

Why not? For similar applications it would be really beneficial to not have to reimplement all the little UI interactions they've polished over the years

Polished is not the word I reach for when I think of Blender's UI

This is just repeating a common meme. Blender's UI layout and organisation is fine, but this is a UI library. There's nothing wrong with blender's UI technically, in fact its pretty decent overall

> Blender's UI layout and organisation is fine

I absolutely detest it (TBF I haven't used it for the past ~2 years) but I also recognize that it's an incredibly difficult problem space that lacks an obviously good solution, it's a free project with extremely limited resources, it has been consistently improving over time, etc, etc.

Everybody hates all serious 3d application interfaces… they’re just complicated. I’m professionally competent in Houdini, Maya, Max, Blender, Zbrush, Marvelous Designer, the UE editor… they all suck in various ways but it’s all a giant set of trade-offs and priorities. Beyond that, lots of these programs have lineages that go back decades longer than almost any modern dev environment and being so complex, UI changes can be quite disruptive. A lot of (no, not all) developers think they have the most complex tooling in the computer world — they definitely don’t.

There's a lot of thought in Blender's UI, and a lot of improvements over the years.

Fundamentally all 3d modelers have inherent degree of complexity. The ones that give a "simple" facade with fewer options make other things harder or having to look into obscure places.

All of that are beside the point however: people will use this as a GUI library, not as a Blender's UX design.

Can you give examples of what are polished UIs in your opinion?

Blender may be the single most polished open source project ever made.

It is actually a miracle that it is as good as it is considering other open source graphics programs like GIMP are horrendous to use.

I'm using it as a video editor for Linux, and I'm constantly surprised how rock solid and fast it is. DaVinci wouldn't even run on my PC without a GPU, but Blender can run on anything I throw at it. And aside from those two, the video editing situation on Linux is a wasteland. I'm so glad Blender added VSE, and that I found it and learned it.

This was true 20 years ago but has not been true for a decade. Blender evolved from among the worst UI/UX on the planet to among the best. It's now fast, intuitive, decently hierarchical and well-organized, leans into search and pie menus, in general it's very polished and in some ways ahead of the curve.

How about a simpler UI consisting of just a text input field for AI.

Despite rumours to the contrary not everything is best interacted with via an LLM. If you want to go to that degree then have at it, all you need is an old school TTY on a serial cable, but the rest of us quite like having UIs designed to fulfil a purpose rather than a soup of Claudeish prose describing a progress bar.

[delayed]

Yeah, nothing like e.g. watching a video or drawing something and changing settings by asking AI.

[delayed]

Egui also has terrible defaults: https://github.com/emilk/egui

But a good design system does the trick. Artcraft apps use this library and their Adobe clones come very close to the original apps.

Every Arftcraft app and non Artcraft related software clones trying to reuse the Artcraft libs (craft-libs) already have everything ready to use.

Yeah but, this can’t possibly be any worse than tkinter.

Q: “Whats wrong with tkinter?!?!”

A: It sucks, that’s what!

Typical HN top comment negativity :)

That UI is fine to be on many more places than Blender, given that many libraries people use are way worse. Imgui for one example.

Definitely a worthy goal but they've gone down a super weird route by providing clones of other GUI toolkit APIs on top of this. Your target audience is people that are ok with vibe coding. They don't care about switching to a proper Blender-style API.

Just because you can...

More like a “why not both” situation, imho. I have apps I could port to use this immediately.

Use the native widget toolkit if the platform you target; your users will love you more.

Most native widget toolkits don't actually support rendering into 3d canvasses very well, nor very compact UI layouts. I generally prefer native widgets, but for a dense prosumer application, they aren't always the correct choice (and you'll see even Apple stray from their native UI widgets for, say, Logic Pro)

They actually do, just not on GNU/Linux, **BSD worlds.

Apple bought Logic.

> Apple bought Logic.

A quarter-century ago, yes. If they haven't migrated to native widgets in 25 years, that might be an indicator of something

It took them close to a decade to migrate the Finder to Cocoa.

I target all platforms, and so does most other desktop software today.

Ok for linux it's GTK and Qt. What's native for windows in 2026? It's a complete joke of system UI.

In 2026 it’s WinUI. Microsoft has been clearer about it: https://learn.microsoft.com/en-us/windows/apps/winui/winui3

But the premise is silly: if Linux has two native toolkits, how is this a sensible recommendation?? Hint: it isn’t, because GTK/Qt are not the native, but rather simply the most used.

Also, go to Mac, and SwiftUI still doesn’t support all that AppKit used to, leaving an incentive for developers to deviate from the platform’s recommendation.

And back to Windows, WinUI still doesn’t fully replace Win32 or Windows Forms, so same conclusion applies.

> But the premise is silly: if Linux has two native toolkits, how is this a sensible recommendation?? Hint: it isn’t, because GTK/Qt are not the native, but rather simply the most used.

I’m going to hazard the claim that in most people’s conception of the term, GTK is native in GNOME, Cinnamon, Xfce, etc. and Qt is native in KDE, LXQt and Sailfish OS.

I am pretty sure HN is an enclave of the only remaining people in the world who think this way. A sanctuary preserving the last specimens of an endangered species.

None of the apps on my computer have any sort of cohesion with each other. It might just be me but I think this idea might be dead, on Windows and Linux at least.

License of GPUI is Apache and BlenderUI is GPL, which is limiting for adoption.

GPL is the best opensource license. But seeing this as a library, it might need to be LGPL rather than pure GPL (ala, you can use it as a library without "tainting" your source, but if you make modifications to the library, you must release it).

AGPL if you wanna be serious about it.

May be, but AGPL is intended for server based apps, rather than local apps (which i presume this library is intended, but i suppose the javascript part can be considered to fall into the server category).

Hello, creator here.

I don't think we can use LGPL since we were derived from Blender's Source code, which is GPL.

ah you're absolutely right; hadn't considered that!

Missed this - this is a deal breaker for many (most?) potential applications.

[delayed]

and thats fine. without LLMs no human would have invested so much required time and effort to do this! and now such awesome things are possible with a bit of tokens and a small prompt. overall a win

Until the project gets abondoned after a few days, perhaps months. :)

Maybe, or maybe not. Let the fittest survive. Also what is developed in a few days can be easily picked up in a few days.

It would be nice if the original chat history with claude is included in the repo. That way design choices are clear and its easier to review and improve upon.

And that is fine; until an malicious actor pushes a PR of "add a botnet that connects to an IRC CnC" somewhere in the code.

It looks good, but the GPL license restricts the use cases

Well.. Old discussion but I actually think it's the better way (sorry. MIT, you are more free then anything). Either contribute back or stay away from open source work that others did for you.

Can you even copyright LLM output? Even the readme points out humans didn't make any of the code.

In this case you just can't spin it any other way than it being a derivative work of Blender. And this is great.

My understanding is that the LLM stripped away the non-UI code from Blender. The majority of the code is ported directly from GPL'd code from the Blender code base.

This is not the spirit of open source.

Recognize that most software is not GPL. The bottom line is when comparing this to alternatives like EGUI and GPUI for building a new app, this no longer can be used, while the others can.

It depends on what you're building. If your app is free software, a GPL library is no problem at all. (and your own code can even stay MIT or Apache.)

It "cannot be considered" only for proprietary apps.

I was just a kid in 2002, so couldn’t contribute, but I remember checking the MoneyMeter going up every day, so that the community could buy out Blender and make it free software. GPL does restrict it, but it’s also only the only reason Blender is not a long forgotten piece of software from the 90s.

Three things one must never ask: a woman her weight, a man his salary, and an HN commenter what “use case” the GPL ruined for them.

In this case LGPL would be a more expected license for UI libraries, at least.

But that wouldn't work for a straight derivative of a GPL project.

in what way?

I believe you could only use this library if the whole resulting app is also GPL.

You can work around that. If you have an application that also has a command-line option, then you could create a UI for that application with this library. The UI module for your application will have to be under GPL because this library is licensed under GPL. But your core program can still be closed-source, as long as you distribute both parts separately. A bit convoluted, but I've seen some applications that do this - they don't distribute the GPL code directly, but ask you to download it separately. As far as my understanding of the license goes, this conforms to the GPL only if (1) the GPL code isn't essential or a core part of the application, and (2) you don't distribute both as one application (which implies that the GPL part is absolutely necessary part of your application, thus violating the first rule).

Is it bui's author decision or a consequence of Blender itself being GPL?

I’ve thought many times “Man, those widgets look terrible!”

I’ve not once thought “Man, those widgets are slow!”

I never found them terrible, but I guess I spent a long time staring at them so maybe I just got used to them.

Do Maya and others have better UIs? Honestly, I find them more repulsive than Blender's

It's obviously vibe coded, but it's quite well done by the look of the ideas and implementation.

What makes you say that? It was the part that goes...

> The repository is built entirely using LLMs, with no human-written code. The drawing and interaction code is ported from Blender's sources, and the behavior is checked by tests: C, Python, Rust and JavaScript suites, the tkinter side-by-side programs and the GPU parity test.

...wasn't it?

not a single mention of a11y or accessibility, so please be careful before using it for general purpose applications.

It'll be added soon!

That was my first thought as well. It would have been reassuring to at least see it on the roadmap.

It wasn't mentioned at the time of writing the comment.

This is excellent. I thought about this years ago for an OpenGL based desktop UI. Apple's UI is using something similar I believe. And the Windows UI too is mostly rendered using Direct3D (which is comparable to OpenGL) since Vista.

I hope someone else uses this for a new Linux distro.

Well, given that so many folks always argue they are Linux distros even though it is an implementation detail for userspace APIs, ChromeOS, WebOS and Android.

The readme suggests it's a port of the blender code, at least the renderer is claimed to be so:

> The drawing code is a C port of Blender's own widget code

How can we guarantee BUI is still performant?

I have a python "DSL" which allows to create a simple UI for synth parameters you also make in python. It uses Tk and looks like, mmm, Tk with all other Tk issues. Project was abandoned because of it. Looks like it's time to revive!

Hi, the creator here.

It's time to do it! Also we have a discord, if you end up reviving it, do share it over here:

https://discord.gg/MWXZfdyM3D

I gave it a quick try (Fedora 44/Python/pyenv) with some Tkinter code which is supposed to just work... It doesn't.

After further exploration, it does work, but not without fiddling around with the tkinter code that exists. Not bad results; better than the original Tkinter rendering, anyway, unless you like that retro look.

Can you please tell me exactly what happened, so I could fix it? Thank you!

Vibe coded slop does not actually work, what a surprise

This project would benefit from a longer Readme.

I have been remaking a Blender style UI in Kotlin for my Umamo project because I love Blender's UI that much.

I think I will keep going with my own efforts, but this looks nice!

Yess, never stop building!

We also have a kotlin binding support coming soon, btw!

It'd be awesome if kitty[1] and calibre[2] adopted the Blender UI.

[1] https://github.com/kovidgoyal/kitty

[2] https://calibre-ebook.com/

Why should kitty use Blender UI? It's just a terminal emulator with no GUI elements and is already GPU accelerated.

jasbr

habs

aibi

I know nothing about anything, certainly not about react, but

    bun add react @blender-ui/react @blender-ui/native
    bun x bui build app.jsx

    import { render } from '@blender-ui/react';
    render(<window title="Is this real?"><label>Hello World</label></window>);

gives me a 7mb executable?! I'm in shock. Maybe it's because I don't know a lot, but this is the most simple way to get from nothing to a GUI I have ever seen. This is, if not the dream, certainly a dream of mine.

For anyone looking at Rust GUI: don't expect GPUI to be the fastest. From what I've heard, Iced, Slint, Dioxus Native, Freya, etc are all faster. Despite claims of being blazingly fast, GPUI isn't that optimized. Zed gets away with it because it's UI is simple.

Really? From what I've heard, Iced, Slint, Dioxus Native, Freya, etc are all slower.

If only there were some way to determine speed based on observable data. Oh well maybe some day.

[delayed]

Why the snarky comment? We already have to data points: one person heard it was faster, another person heard it was slower. If more and more people tell what they've heard then we will know the truth. Please go read how science works!

Now if only we could peer review our peers’ reviews..

Some actual numbers (vibe coded benchmark ported from the JS framework benchmark):

  Median update+build time per frame (ms), lower is better
  Machine: macOS 26 VM, Apple M4 (virtual)

  benchmark (ms)      Dioxus    Iced    egui   Slint    GPUI
  create 1k            12.91   17.57   10.66   26.99   22.32
  replace 1k           14.46   14.48    7.48   17.07   17.41
  update every 10th     2.12    1.94    4.17    5.55   10.65
  select row            1.16    0.37    3.50    6.62   10.68
  swap rows             1.45    0.42    3.53    4.10   10.26
  remove row            1.32    0.39    3.27    4.46   11.95
  create 10k          124.97  152.52   58.97  136.07  164.58
  append 1k to 10k     28.12  157.67   33.01   69.93  117.70
  clear 1k              3.00    0.86    0.07    4.60    0.04
  scroll frame          0.28    0.10    0.55    4.23    3.79

Dioxus above is Dioxus Native.

These are CPU-only times for the state management / style / layout / etc, but don't include rasterization (additional work in the benchmark harness required to measure GPU time sensibly).

Did you test gpui_fast? I'm curious how it compares.

I did not, but judging from it's README it may well solve most of the issues.

Do you have a repo or something i could run the benchmark locally with please?

This must be the lowest effort hacker news comment ever. And I've made some really low effort myself.

It is enough to counter the original claims though.

This matches my findings. Asides from some performance footgun, it all strikes me as a single purpose, low level thing with a DX only an LLM could love. I suppose many today don't need more than that but I found the other Rust GUI solutions to be maturing rapidly with better stories across the board.

The main thing GPUI has going for it is mind share and that's a great thing to have. The mind share is unfortunately highly fragmented though as the original authors are not interested in selling a GUI library. Hopefully the community efforts will come together to produce a more polished and general library. A well used library is better than a better designed one maintained by a single person.

It really annoys me the rust library is named “GPUI”. That’s the name of this method that’s been used for eternity by games.

It’d be like naming my http library HTTP.

And it’s why Dear IMGUI isn’t just named “IMGUI”.

OK who's gonna burn the tokens rewriting a whole Linux system with BUI? You know you want to...

Lol, I'll showcase that project in the website if someone actually does that

BlenderOS

that's gonna be crazy. not gonna lie, that's actually doable

re. "macOS written and waiting for its first real Mac" ... Apple give this man some hardware.

wish apple was that generous :)

I converted a realtime robotics simulation platform I've recently vibecoded in this thread's honor and it worked great first build.

Do it!

"Decoupled from project by external developer" generally means "with a very high chance of immediate bitrot".

If you're going to ask Claude to write a library with a web version, can you also ask Claude to write a web playground?

> Every chapter links to a longer guide in docs/ or on the website when there's more to say.

My slop detector is tingling

I've been building a desktop app on GPUI on Windows, so a data point on start-up: most of our time to the first frame wasn't GPUI's layout or text at all, it was creating the Direct3D 11 device, about 150 ms, which is the graphics driver loading. Moving the device creation onto its own thread at the very start of the Windows platform setup, so it runs while everything else gets ready, took us from 278 ms to 222 ms. Memory at idle is similar: most of it is the driver, not the UI. So on Windows I'd compare these frameworks with the GPU setup measured separately, otherwise the driver dominates the numbers.

Nice! I've been waiting for someone to do this for a long time. Now I'll just need to wait a bit longer for someone to apply this to FreeCAD. Boy oh boy!

That's really cool! I burned some tokens trying to develop something similar but gave up halfway. A lot of people criticize Blender's UI but as someone who has been using blender for more than a decade now I just find it so elegant. Just the customizability itself is something I miss on most 3D software, if not all of them. Being able to just split screens and change their functionality, paired with the tabs for different workflows, it is a really good solution for a very complex software. I'd like to see a game engine with this much UI freedom.

Do you think it's with trying to revive Blender Game Engine? Edit: just learned there's a fork https://upbge.org/

BGE underperformed in my machine a lot, since AMD Athlon days. Godot it's far better.

I love the blender game engine. To me, it's as nostalgic as a favorite video game from childhood. I spent countless hours building little projects as a kid. I was really sad when they removed it in v2.8

Now as an adult, I can see why they removed it. It didn't make sense to put resources into it.

I'm really happy about the UPBGE project. It doesn't compare to the capabilities of Godot, but that's okay. It just makes me happy that someone else cared about it enough to pour time into it.

There's also armory: https://armory3d.org/engine/ / https://github.com/armory3d/armory which is the same idea I think

Hi, the creator of BUI here.

I do hate vibe coding. So I took a different approach for BUI.

Instead of asking the agent to decouple the UI, I asked it to remove all the parts which isn't a part of UI logic.

Turns out AI is pretty good at removing than generating.

"It isn't vibecoded, it's vibe-pruned!"

That's an interesting distinction. A distinction without a difference in my opinion, but a distinction nonetheless!

lmao, this made me laugh

I still think Softimage|XSI was the best UI/UX that ever existed. Moving to Blender from it was... painful.

Be aware of the license. You can only use it in GPLv3 programs.

Nice. I have my own GPU GUI toolkit I'm satisfied with and developed because I did not want to have yet another imgui looking app that required gpu accelerated UI toolkit.

But, if I did not invest last two years into my own toolkit, I'd definitively consider.

I'm interested to learn more, do you have link?

Cool but Blender's interface is not a general purpose UI. EGUI/GPUI and the rest are.

Others have noted that GPLv3 is required (which is a show stopper for many) but I'd like to mention that I find the pill saying "GPLv2 or Later" in the project itself misleading.

Especially since v2 and v3 have vastly different consequences.

Can I please congratulate you on your README. It rapidly tells me what I'm looking at, and why I might be interested in the project. There are so many READMEs around that ... just don't do that. They ramble about technical implementation details, how their language choice is amazing, and a third of the way through the README, I'm still left wondering, "but what does this project do?".

Yes, this is so important and yet so often not considered. This README was pretty good, thorough and not ambiguous. The number of projects that assume you know what their obscure idea relates to is huge, mainly as the authors can't put themselves in the shoes of their audience.

Yes, this is definitely one of my pet peeves. Just give me input --> output, then everything in between.

Many of my old >1K star GitHub projects were not that great, but I always made sure to create the cleanest, simplest possible READMEs, inspired by things like the Python requests library's docs (https://github.com/psf/requests). I think it helped a lot in their success.

I remember someone once created a GitHub issue requesting I turn the README into some verbose technical thing with full API docs and a bunch of people replied saying they greatly preferred the "pitch + high-level examples" style.

Would be good if you could write the README without an LLM

I can't figure out if you're being sarcastic

The very first line of their README:

"Blender's user interface, as a library you can use from Python, Rust, JavaScript (React, Svelte), C, C++ and Dear ImGui."

This tells me a huge amount of what I need to know, if I'm interested in this project or not.

thank you, I thought you didn't like it lol. let me know if you use it.

We also have a discord server, lmk if you're planning on using BUI!

[delayed]

It is complete, if you realize that its the subtitle to the "BUI" title. Read it as if there is a colon between the two.

isn't dear imgui a gui library already?

It still sounds like LLM which is nauseating to read. For instance: "Blender's interface is famous for how it feels. You press a number field and drag: the cursor vanishes, the value glides under your hand, and you never hit the edge of the screen. " that's ChatGPT writing.

It is ALL AI generated I can say that because I recently worked on a landing page and wrestled with Claude Opus the WHOLE way through. Claude has no idea how to write copy. Neither do any of the others, you have to edit with a fine-toothed comb.

To the project maintainer: don’t ask people to read a terrible AI generated sales pitch novella that quantifies your AI generated port against every UI framework out there.

Here’s the obvious giveaways:

“ 4. The many ways in” “ Python, four ways”

“ tkinter: all 7,846 public names of tkinter and ttk. Nine real programs behave identically on the standard library and on BUI.”

“ Rust: the web platform, natively”

Ugh it’s excruciating. You’re better off with a readme that shows a few screenshots and installation instructions.

[delayed]

My mind checks out when I see LLM prose. Didn't get that far.

[delayed]

I could handle the API being LLM generated (but maybe gently prize it out over time), it's the copy that gives me LLM nausea[1]

1 - thanks to the poster that mentioned nausea, I've been looking for the word to describe LLM based writing for a while and it's perfect.

[delayed]

i am way too lazy to write good documentation. BUt i also dont want to have anyone read claudette-prose so i noticed that

"down to earth" and "hobby-project" does a good job. its still not perfect but better than the vanilla output. eg. https://github.com/busssard/OpenRing

it's funny how these claudese constructions also come out of lil baby local qwens and stuff too. i guess they're just directly distilling a claude endpoint, or whatever.

like, if people are gonna be turning the crank to generate prose-like text from statistical models and calling it a day, aren't "subagents" a hip new thing nowadays? there do exist models that aren't overcooked to impressively bad long-tail in the pursuit of maximum coding benchmark scores. such other models generate prose-like text that is significantly less miserable to wade through.

if you're a vibecoder who's terminally committed to the bit, for god's sake, at least tell a local gemma to draft your readme or something. an occasional falsehood is less damaging to usability than the experience of reading feeling like trudging through a pool filled with treacle.

This is why I stopped using Claude. ChatGPT is better in that regards, but you still have to write a ton of rules for it to read half-decent and then edit it manually anyway.

None of them is better than the others on the frontier. But to be honest, the models are designed for acing STEM exams not doing copywriting.

This is further evidence of the need for small models doing specific things well.

Where I would say models fit in the process is establishing the general layout of documentation and iteratively refining things out. Interestingly, it’s the same for image generation too. But 9 out of 10 it creates incomplete work that needs refinement.

Go wide, then hone in, refine. There’s no such thing as getting AI to single shot creative tasks. Which is funny because everyone thought writers are on the chopping block initially, but that is unequivocally NOT the case.

Even technical tasks, agents doing it their own way yields garbage in some right. It’s just a question of how long it runs before it goes down a path of no return.

Human in the loop is the only way. And, to be honest, what’s the f*king rush anyway? We all need to slow down. Good ideas with unravel in time. :)

Thanks for the word nausea - I've been trying to put my finger on that - I kept thinking of sickness but that's too much, LLM nausea is definitely a thing.

Not to mention that it includes screengrabs! Which a shocking amount of gui and web projects dont do

There's some effort spent into it, unsure how much of it is LLM. I find the readme very confusing. It spends a lot of time talking about Qt APIs, and dearimgui... most of which could go in a separate page.

But the worst part is all code snippets are followed by an image which is not related to the code snippet. It's not what you get by typing the code... it's something completely different.

Did you skip Part I?

I'm confused. Are you being sarcastic? Isn't this readme in the most typical LLM-ism?

[delayed]

...really? for my part, after a few paragraphs i felt in my bones that i was reading undisclosed llm output and noped out of evaluating the entire thing any further.

but also the fact that yours is the most upvoted comment tells me there are entire swathes of people for whom that style works, and i am interested to get in people's heads about that, because it might yield some insight re: for what kind of audience these labs are optimising the style of text these models generate. so, if someone is reading this comment, and you broadly agree with the comment i am replying to, please answer these questions:

1. is english your native language?

2. what about the prose do you find appealing?

3. do you not find the structure (colossal wall of text, infodumping about largely irrelevant details) annoying/noisy?

4. does the word-to-word style and clichéd phrasing not immediately set off your llm alarm bells?

i guess people can answer if they disagree too, if they really want to

I could tell within a few sentences that it was naked slop, and I was grateful to be able to nope out rapidly, without missing anything.

I liked the README, or what I read of it.

1. English is my native language.

2. The organization was to the point (eg it described what this is and who would use it up front). I can't say I found the prose "appealing" -- about the best I can say is that it was so obviously LLM writing that it implicitly told me that this is most likely a large mostly unattended porting project, which helped me set my expectations accordingly. Note that this is also why I stopped after getting a good feel for what it all is; the project would be interesting if I had a need to make something with that exact style of UI, but otherwise there wasn't going to be some shiny new idea or brilliant perspective that I could apply more widely.

3. Partly. It was a wall, but not that impenetrable a wall. There's plenty worse, and it was broken up enough for me to get through. Things were signposted well enough for me to skip a lot of what wasn't relevant to me.

4. Oh, absolutely. See above.

I'm finally growing the ability to extract value even after detecting that text is in the en-SLOP locale. Originally I couldn't detect it, then I could and couldn't tolerate it, now I'm starting to be able to treat it like broken English or caveman speak. I have to work to extract value, albeit a different type of work than those two.

It's mostly about suppressing my reactivity to the choppy phrasing and braggadocio (which are both the results of reinforcement training to elicit exactly the sort of reaction that I'm suppressing).

I don't get this comment, the readme felt like slop to me. What makes this any better.. I genuinely want to understand why someone thinks this is a good readme sure I may have seen worse but I have seen much better readmes.

The very first line of their README: "Blender's user interface, as a library you can use from Python, Rust, JavaScript (React, Svelte), C, C++ and Dear ImGui."

Then, the README very quickly continues with "What BUI is" and "Why BUI exists". These are all the exact things I want to know, ASAP, when I encounter a new project. The language reads perfectly fine to me.

Here is another example of a nice README: https://github.com/mpv-player/mpv

Do you know what I don't want to know first? What the startup time is, what language you wrote it in (because you think it makes your project cooler), some other random internal details about the performance of the thing, when I don't even yet know what the thing is.

> what language you wrote it in

You mean something like:

> turns it into a standalone UI library with a small C core

...?

Or this:

> At the heart of BUI is the kit: a small runtime written in C

......??

I personally have no issue of this, but the more I read your comments the more confused I got.

I will uncongratulate the README. Linking back to the original HN post (and including point count!!) is peak "I'm doing this for internet points".

Blender continues being the GOAT

No system theme matching? No system fonts? A fast UI toolkit is great, but so is one that matches the rest of my desktop.

Is this even a thing anymore? The only app I can think of that matches the system theme is GIMP… not exactly a paragon of design

Uh, all the major browsers at least attempt some theme-following.

creator here, will add it soon!

I have the same idea 10 years ago. :D I am so happy seeing that somebody was capable of making it. I always admire blenders UI and all the work developers did on making it.

I don’t buy arguments like much faster than X. Take X let claude work on it for 2 days and it will be fast. In case of gpui I took the community version as a part of my project and it is quite fast because it is optimized for my app.

so, why gpui not taking Claude?

But is there a UI editor?

Is it retained mode UI?

But, can it work without GPU? All native toolkits have fallback to software rendering. I do not uderstand what is gained, cpu was good enough for that since Windows 95 era? Just do not redraw every frame like in games but retain instead?

Yes, it defaults to CPU and explains when you might want to use the GPU and why in this section https://github.com/ppmpreetham/BUI#18-rendering-on-the-cpu-a...

The project is dishonest. Maybe AI agent was instructed to split Blender UI code, but what happened instead is AI implemented a classic immediate UI, not even reading the code.

I just tried few very standard operations that every Blender user knows:

1) Drag vertically across multiple number inputs to enter multiedit mode - failed (even though main page advertises this: "Drag down across X, Y and Z to edit all three at once")

2) Hover on field, press Ctrl+C, hover on another field - press Ctrl+V - failed

3) Press shift while dragging - it glitched

4) Scale UI - failed

5) Collapse collapsible block - failed

6) Reset to default - failed

7) Jump to next field with Tab (while editing) - failed

So basically it is more like styled imgui (less polished than Blender), rather than Blender's UI.

Hi, the creator here.

It works perfectly on my machine, and another friend's machine. Just to be sure, which demo did you try? the website or python or rust?

I checked only web samples in https://bui.preetham.top/

Probably it works in other implementations (I did not try); in web it produces <input type="text"> and agent likely gave up trying to simulate such details there.

Oh yes, the web implementations there for mockup.

Let me update the web samples. Thank you!

they're updated, check it !

You completely missed what I wrote in my message. Web version is still neither a original decoupled Blender UI logic, nor an attempt to reimplement it in js. Yes, now when you move cursor across 3 inputs, value changes, but that's not even close to what original Blender controls do.

If you want share a demo, maybe canvas-based would work better (e. g. https://www.egui.rs/)?

I'm excited, and will try this! Its provenance speaks loads, as Blender is a beacon of software design, open-source or otherwise.

I'm an EGUI fanboy, but am looking to truly compare it to other things, especially non-immediate mode. I have no gripes with EGUI, other than it seems to create an application size floor, and I suspect memory use could be better. It has never blocked me, including for complex applications with GPU rendering etc.

Blender is peak OSS. I'm in! Will give this a shot for my next limited-scope application. I'm also experimenting with EGUI.

I will try Slint if I ever do embedded with a full GUI.

This looks like a nice quick option for python scripts. I've made a web interface using fastapi for my python script because that seemed the most sane choice at the time but I would've used tkinter but for my lack of familiarity with tk/tcl and it's retro look.

Yessss, I'm using it myself for my quick and dirty uni projects currently.

Also speaking of FastAPI, have you checked out FastRAPI?

Amazing Amazing Project

Love Blender, but it’s GUI, especially anything that could use native OS components, is not what I love about it (judged on macOS).

Hi, we currently don't have a macOS build yet, we're planning to do both native and GUI both !!

But is it built in Rust?

The source code is in C/C++, there's no need to port it to rust

> The GPU renderer is opt-in: Direct3D 11 on Windows (gpu/intern/gpu_d3d11.c) and OpenGL 3.3 core on Linux (gpu/intern/gpu_gl.c).

Direct3D is at version 12 on Windows[1].

What tech on macOS. Don't tell me OpenGL.

[1] https://en.wikipedia.org/wiki/Direct3D

Those numbers look very impressive, I'll try it in some projects

Why should I consider this over say PySide6 for Python apps?

That README is the most uncanny thing I've seen recently. I still have no idea what this thing actually is, what the layers are if any... it's talking about C, Rust, HTML, CSS, node, bun, react, Python, Qt, QuickJS, DearIMGUI, etc.

But what is actually in that project? All of those things? Just a C runtime? Why not talk about what the barebones C runtime gives you first (memory, runtime), then add whatever layer on top? Or is it all-in-one always?

It's just AI slop

This looks great, are the docs mostly complete?

Was recently trying to learn a desktop toolkit again and was disappointed that the docs of pygtk and rust/iced end around page three. Many projects, little follow up. Probably should pick a toolkit from the 90s instead.

This seems different. Also, tabs that look like tabs! Love it.

Yes, we made sure that you wouldn't need docs, you can write your code in your favourite framework, and it works properly!

Check out the docs here: https://bui.preetham.top/

>People spent decades building and optimizing Blender's GUI. It's battery and memory efficient, and super duper fast. We decoupled it, to give everyone a chance to build their own products with it.

Who is "we," and what's the long term plan?

It's a shiny presentation, but I don't see anything on offer more than a solo dev who pointed AI at the Blender codebase. I could do that myself, and many of the reasons I don't do that apply directly to this project too.

Curious what’s the plan to keep it in sync with blender changes, if any? GUI layer doesn’t change often but still

currently it looks great! we can always keep adding the newer components and other stuff

This is cool. But can I write the UI in an XML file?

Give me an example

Look up TouchOSC and its layout engine. I need this functionality in BUI

"People spent decades building and optimizing Blender's GUI. It's battery and memory efficient, and super duper fast. We decoupled it, to give everyone a chance to build their own products with it."

Is it just me, or is this late stage token-utopia derangement? It took me quite a few tries before I could put Blender to any actual use because of how ... unusual ... the UI was.

Feels like intentional brand-confusion with Blender.

It is a derivative work, isn’t that appropriate? Or are you saying it shouldn’t have blender in the name unless it’s owned or officially sanctioned by the company that owns the blender trademark?

It's just obviously trying to ride coattails with vibe-shit. If "blender" wasn't in the name it wouldn't be on the HN frontpage.

I have been using blender since 1998. In the mid 2000s I pulled out the blender UI for a project but never published it and the project died. I have been waiting for a project like this for 28 years! Amazing job.

Question, is there an effort to have blender use the library? Avoiding drift would be really important.

Yea, but UI/UX looks very bad.

Why does the Rust example look so much different than Blender? Or the other examples?

Why are the progress bars gradients? Blender doesn't use a single foreground gradient. Why is there a shadow on the thick status bars in the middle? Why are headings different colors? Why are there badges? The fuck is this, tailwind? The icon on the dropdown menu top right isn't centered properly. Right to jail, right away

Love this.

Most bottlenecks are not part of the framework, they are part of your application.

I'm really liking https://github.com/anomalyco/rune so far. Designs come out much nicer looking

I added Lua binding support for anyone looking for it, just got merged upstream ..