Adobe Creative Suite Cleanroom Port to Rust | Hacker News Reader

Adobe Creative Suite Cleanroom Port to Rust

github.com

144 pointsbigwheels91 comments

Open articleView on HN

https://github.com/storytold/photocraft (Photoshop)

https://github.com/storytold/vectorcraft (Illustrator)

https://github.com/storytold/filmcraft (Premiere)

https://github.com/storytold/lightcraft (Lightroom)

https://github.com/storytold/printcraft (Acrobat Pro)

https://github.com/storytold/effectcraft (After Effects)

https://github.com/storytold/designcraft (InDesign)

I love these apps! The author is on HN too, we were discussing Rust GUIs the other day:

https://news.ycombinator.com/item?id=49922904

Oh nice, I didn't realize it was echelon.

What is interesting is how fast this got flagged off the homepage, it would be really nice if people associated with these companies disclosed their affiliation and would refrain from flagging this sort of post.

Kind of seems like the thing the mods should be on top of, doesn't it? Something getting flagged because users have a conflict of interest and a pro-corporate bias.

Interesting but for photography at least, there's a massive hole: RAW support. What they've implemented is very limited and basic, nothing like parity with Lightroom or Photoshop. Just try loading a RAW photo into Lightroom/Photoshop and put the window next to Lightcraft/Photocraft - the results are completely different and not in a good way. Lightcraft's image looks like an amateur attempt at RAW processing from first principles, without any of the most basic colour profiles and adjustments that Lightroom applies. Photocraft doesn't even offer RAW processing, it just dumps the raster straight in.

I'd treat this as completely new, far from complete software heavily inspired by the Adobe Suite, rather than a "port".

But maybe folks who use workflows that don't involve RAW photos have better luck.

Or maybe you'll have better luck in a week when somebody familiar with RAW and a 20 year grudge against Adobe chooses to spend a day of tokens on the problem. Maybe me.

Rome wasn't built in a day, nor Adobe destroyed in a single git push. But they both shall fall.

That's exactly what I was thinking. It's only a matter of time. Surely the models, when well directed, can get most or all of the way there in short order.

Surely the models, when well directed, can get most or all of the way there in short order.

"Hey ChatGPT, what would the sufficiently smart compiler argument look like in the LLM era?"

I did this a few months ago and it's not simple: different sensors have different pixel arrangements, files can have different data structures, and the default tone mapping is provided by the app (so opening the same raw file in darktable vs preview will have different results). There are different algorithms for how you assemble the raw data as well, and the issue is that you can get something _passable_ that isn't _optimal_, and it's really hard for the LLM to recognize that something isn't right. Is it the pixel processing, or is it the tone mapping, or is that just how the image looked?

And on top of that, you need a bunch of different files from different sensors to test with. Canon, sony, apple, fuji, nikon, olympus, etc. -- if you're building a truly general application, that's a concern.

Or if you're like me and you shoot fuji, you just build a fuji raw parser and call it good. The "TAM of 1" world of software development...

All the camera companies offer custom software that works better (mathematically) than the Adobe versions. They're just trapped in godawful UX. It's a one day job to do the brain transplant.

Adobe employee here. We have access to LLMs and tokens as well.

And 500 levels of org tree, and grumpy customers that freak out if you change anything about their workflows, and decades of legacy code across dozens of micro-services, and backwards compatibility requirements with 20 year old formats...

You have:

1. LLMs

2. an executive team that hates your customers. this problem with company culture goes back at least 13 years: https://fstoppers.com/video/adobes-ceo-deftly-dodges-questio...

Unfortunately any benefit from #1 is neutralized by #2

How are LLMs and tokens going to help you retain customers who you've spent the past decade pissing off?

Those customers are already pissed off, yet record sales every quarter. This will be great to offer this diminished experience to them until they start using AI to do all of their editing.

I'm going to revel coming back to this arrogant comment in a couple years.

Seriously, you have no idea the kind of goodwill you've squandered, and the kind of resentment you've created. I expect it will bite your company squarely in the ass.

I was surprised as well that an Adobe employee would make such a revealing comment.

Call it arrogance or just an acknowledgement of reality:

Revenue over the last ten years

August 31, 2026 — $6.760B May 31, 2026 — $6.618B February 28, 2026 — $6.398B November 30, 2025 — $6.194B

August 31, 2025 — $5.988B May 31, 2025 — $5.873B February 28, 2025 — $5.714B November 30, 2024 — $5.606B

August 31, 2024 — $5.408B May 31, 2024 — $5.309B February 29, 2024 — $5.182B November 30, 2023 — $5.048B

August 31, 2023 — $4.890B May 31, 2023 — $4.816B February 28, 2023 — $4.655B November 30, 2022 — $4.525B

August 31, 2022 — $4.433B May 31, 2022 — $4.386B February 28, 2022 — $4.262B November 30, 2021 — $4.110B

August 31, 2021 — $3.935B May 31, 2021 — $3.835B February 28, 2021 — $3.905B November 30, 2020 — $3.424B

August 31, 2020 — $3.225B May 31, 2020 — $3.128B February 29, 2020 — $3.091B November 30, 2019 — $2.992B

August 31, 2019 — $2.834B May 31, 2019 — $2.744B February 28, 2019 — $2.601B November 30, 2018 — $2.465B

August 31, 2018 — $2.291B May 31, 2018 — $2.195B February 28, 2018 — $2.079B November 30, 2017 — ~$2.008B

August 31, 2017 — ~$1.844B May 31, 2017 — ~$1.773B February 28, 2017 — ~$1.682B November 30, 2016 — ~$1.611B

And yet... 10 straight quarters of ARR decline. CEO leaving. Market prices you like a company in decline.

Besides, you're on HN, we don't value revenue here because we're idiots. But even the market agrees. Stock down 64% from peak.

Tootsie Roll Industries heats sugar and has nearly triple your P/E.

Huh? ARR has been monotonically increasing. Maybe you mean the ARR growth rate? Yes, that has been decreasing.

You also have managers actively hostile towards your customers.

That is true, but does Adobe have another Photoshop to build? And in how many hours will that be copied as well?

Not too many years ago, you had significant open source C++ libraries. You cared about them as much as you cared for customers.

So, what's Adobe Acrobat's competitive advantage to justify $15-25/mo vs. a free PDFCraft?

I haven't tried PDFCraft, but I'm willing to bet it isn't going to handle tables as well or offer signing.

Anyone would be insane to pay for Adobe pdf over pdf xchange, but plenty of companies do. I'm not sure how/why the world pays for Adobe with so many good options but they do.

People are underestimating heavily what has happened here and to our industry in general.

Glad to see another based knower knowing Rome still exists

I chose "both shall fall" (future tense) on purpose. And likely in the same manner: a self-destructive period of reactive fascism, followed by becoming a nostalgic name noted for its crumbling architecture.

I also built one of these photoshop clones. It has raw support. I use it for Fuji raw editing. I've been building it with Claude for 6 months now.

https://lopsy.art

How many painting applications have been built by LLMs I wonder?

I also let it build one btw: https://egeozcan.github.io/ketchup/ (not a PS clone, more amateur painting oriented)

Yeah, it's a really well defined problem space with quantifiable outcomes, so LLMs are very well suited to it.

For lopsy, the biggest issue I ran into was combinatorial bugs: certain sequences of tools and actions would cause a bug to occur that wasn't obvious from isolated tool testing. I solved it by having Claude generate complete projects, full documents with novel concepts chosen randomly, built in the browser with playwright. This continues to expose so much more than individual unit/tool tests, and I have an automated loop that addresses them. (And a bonus from that: I use the outputs to make tutorials to show what the tool can do).

>amateur attempt at RAW processing from first principles

That tracks.

So it's the Omarchy of Adobe-like software?

Raw was a tough one when I tried it a few months ago -- I never thought about how individual color pixels on the sensor come together to form a color value for a point in an image file. Super interesting, and not at all standard between the different manufacturers. It's a hard problem for software to solve in a general way.

why do you want to write your own? we do have libraw supports basically every camera raw format

I find it awfully funny that your response to someone creating the entire Adobe software suite in a day is “the colours aren’t rich enough for my liking”

Oh the times we live in.

In two of its test cases, it changes the PSD on save so much that an export render differs. No RAW support. No folder sync or plugins. Limited brush support. It’s like the author never actually used their vibe-coded output and just decided it looked good enough.

this sounds exactly what I'd expect from v0.3.0 of a software.

a lot of rough edges of this program I see are actually features I never even used myself in photoshop-like program that much, so if I was doing that kind of AI rewrite I'd also miss polishing them. publicly releasing this right now is exactly the move author needs to get human feedback from people who actually use these features daily

I was able to use Photocraft and Claude to do something that the plain generative image AI in ChatGPT couldn't. I had two photos of the same location, one with 3 people, one with 1 person. I took the photos with the people strategically standing where the photos could be easily be combined without the people interfering with one another. A month ago, I tried using ChatGPT to combine the people into one photo. Its attempts never looked right. It always changed the faces of everyone. Today, I tried using Claude and this Photocraft MCP server to do this. After 3 quick tries over 5 minutes, it did it. The 2 additional tries were to have it fix some lighting differences. Thanks to the person who created this!

Yes, I could have definitely done it myself, but that's besides the point. It wasn't important/urgent enough for me to spend 20-30 minutes to do it (yet). But I was curious enough to play with this to have it done this way.

Post AI - What is the future of making money with software? Desktop apps or SaaS? This certainly also would cut into the bottom line of software companies so they won't exist and then these companies are supposed to be the prime customers of the AI companies.

Can't understand what's happening.

PS: Anyone badmouthing the above rewrite is not understanding what's happening. Check back a year later for featur parity.

I wonder if software will be built inside of companies that have non-tech competencies going forward. The story of 2000-2025 was every traditional company becoming a tech company. The next 25 years might see tech companies without any offering besides "tech" being at a disadvantage because their entire offering is easily replicated.

In this hypothetical future, why would I (or my business) buy software from any company, when I can just prompt the equivalent into existence myself? We talk a lot about AI killing software engineering as a profession, but if the AI future comes, it will kill software development as a type of business.

Read the git log from this software and it will be very clear very quickly. This can only be done with careful guidance from experienced developers. It doesn't work at all for a vibe coder.

Maybe that'll change I guess, but this repo is very bad evidence for that.

Nobody is disputing that this requires software engineering. Look at their architecture alone for the evidence.

We're talking about employment here.

Because there are only 24 hours in a day, and you have other problems you want to solve. You don't have time to design a car from scratch, you just buy one because you need one to get to work. At most you buy a cellphone holder from the gas station or that one you saw on TikTok. And a bumper sticker. Yeah, if that software is bread and butter for your business then sure. Go right ahead. But business is about delegating. To employees and to vendors. It comes down to data interchange. What does your business have as inputs, and what does it have has outputs? For some, it's emails with PDF attachments, others it's an API accessible via http. Others it's raw ingredients and the output is food. Sometimes you go to a restaurant and pay someone else to do the cooking. Other times, you cook yourself.

I give you 2 bucks for using that vibe coded tangled mess i percieve as working?

Previous thread:

https://news.ycombinator.com/item?id=49958850

One of the highlights in this for me is that we are starting to see clean room ports of software that companies would actually sue to protect.

Clearly the old "nobody would re-implement this" protection no longer works. I think we may stand on the brink of software losing a lot of it's magical DMCA protections, because it no longer matters. Or worse, the law becomes stricter.

What DMCA protections could Adobe use against this? Surely the look-and-feel case established they can't trademark their signature keyboard shortcuts?

can't they say you broke the DRM? and sue? like for example same game companies even sue you for touching the running memory of the app

They'll come up with DMCA II when this start being a big enough problem for them, don't you worry.

There are several of these clones floating around now. Who is furthest along in functionality and community support? Have any of these attracted a community of contributors?

I think this would be the defacto open source alternative to Adobe for the following reasons:

1. it's a rust based desktop app and not based on electron.

2. it's cross platform - available on win, mac, linux and wasm.

3. it's a suite of apps and not just a photoshop clone.

4. it's free and open source.

5. it looks good.

6. it's amassing a community of contributors.

There are other photoshop clones around but none with the ambition and scale of this project.

Also it got a lot of press.

Wow! This actually gives a pretty promising alternative to GIMP on Linux. I was resigned in my fate to just deal with the awkwardness of GIMP and Inkscape after years of the creative cloud suite. It is clearly missing much at the moment, but if they can keep going this would be amazing. All packaged as AppImages so no need for the persistent junk running on the system.

The UI is great but even basic features are relatively incomplete. It has potential, I hope it gets to places.

I released something similar today: https://pixellect.app/

Though it was done roughly 1-2 months ago, a very busy work / private life kept me from finally releasing it. Interesting to see this being released on the same day. Though Pixellect has a different mission. Its not a clean room photoshop clone but more between Pixelmator and Photoshop and also features (I think) a bit more AI Integration. However Photocraft looks much more complete in terms of typical Photoshop features.

It was only 4 months ago that AI development was in question because there were no vibe-coded Photoshops. Here they are, and I think more will come.

I've been doing the same thing on lopsy.art.

I wonder if we're entering an era of boundless open-source clones and variants of proprietary software, from operating systems to games, products and services, perhaps even clones of entire companies.

I think so.

I’m putting personal attention and tokens towards replacing Pioneer / Alpha Theta Rekordbox (music management for DJs).

The software is old and clunky, has a ton of features I don’t use, and can’t customize what I do use. And it’s enshittified with ads and popups and an expensive monthly subscription.

One big motivation for me is that it is FUN to rebuilt it with a coding agent. I’m learning a ton and the custom tools actually work.

I know I’m not alone.

https://deadca7.com/

Tower of Babel, but for software.

More like chainsaw for attention pie.

I hope not, that sounds terrible waste of time and effort for creators and consumers alike.

Hasn’t it always been like this? I feel like a few simple options is somehow better than a monolith complicated option.

The speed that the Premier clone launches is mad.

Can't encode 265 or AV1 though.

Tried the Photoshop part. It actually not bad.

That being said, I see at least one massive problem with all those Rust "clean" room <x> <y> and DHH — Rust allows to iterate too fast in somewhat safe fashion. Sounds cool, right? However, that crates a gap between humans and machines (agents) in understanding what's right and wrong which leads to unmaintainable software.

I'd love to see more "clean" room <y> using C/C++/Zig. It would be slower and probably less "safe", but at least humans can keep up.

Let's see where that "echelon" in 18 months ;)

This is sweet. I've already modified it to have a couple features I miss from Krita

1: Create new document from image data on the clipboard

2: Custom grid options (set width AND height of grid cells, offsets, etc)

I stopped using PS a while ago when it stopped working well under Wine. Getting to know Photoshop again after so many years is really fun

Has anybody with deeper knowledge of the originals tested the functionality?

What's the catch?

It's slop, and there's no telling if it will still be maintained a year from now.

I don't think you realize the depth of Adobe creative suite hate, an entire generation of artists came up on Krita/Affinity because Adobe decided to vampire its customers.

It's deep with a small number of very loud people.

...which is one of the benefits of it being open-sourced with a permissive license. If it stops being maintained, it can just be forked.

The "no one maintains AI slop" idiom is more for things which are bespoke and/or don't have much demand, and this clearly has demand.

Why would anyone help maintain this when they could vibe code their own though? Genuine question. As of the time I'm writing this there are already two other people linking their own Adobe clones in this comment section.

It takes a lot of tokens to vibecode something like this with the expected feature sets, and this domain is complex enough with enough edge cases and interactions such that it's infeasible to fully QA even programmatically.

Why ? Obviously it was possible to write Photoshop itself with human QA, and human QA is fundamentally limited ... and is more limited than programmatic QA. At the very least this repo will start to catch up fast with Photoshop, assuming it's maintained.

The website is entirely vibcoded and is thus free of useful information. I haven't tried the applications

It's just slop. Tried to install it via flatpak. It has flatpak instructions in the readme. It does not ship the flatpak files it references in the readme.

It hasn’t had a massive QA team (and customer base) go through all the missing features and bugs.

No doubt it’ll get better in time.

More likely completely neglected.

Missing and incomplete features, missing templates, presets and it's likely to be abandoned as issues will start to pile up.

The user experience is also not great with professional looking AI apps, I've just tried PhotoCraft and couldn't paste an image from the clipboard. Clone stamp doesn't have a live update, etc.

Another catch is that it's... PhotoShop. I'm not particularly a fan of the whole workflow. Creatives learned to live and work with it over the decades, so it works, but it's not really an intuitive tool.

I'd recommend looking into Affinity which is a merged version of PS+Illustrator. The product feels like it at least tries to be a little different. It's also free.

Try lopsy.art. I've been at it for 6 months. Updating and building out features almost daily. It's a Rust/WASM engine too. Also open source. github.com/theseamusjames/lopsy.art.

I think there are about a dozen of these now.

Why is this flagged?

HN is filled with AI haters. They press flag at every positive thing about AI.

This is a more interesting one, though; some people might dislike Adobe's practices enough to be okay with an AI reimplementation of their suite.

I just wish I could still buy a 'Creative Suite' one off and keep using it instead of having to subscribe to these apps.

Besides their dark patterns, whenever a company participates in the shady data-broker industry at Adobe's level, I celebrate any such success in OSS.

AI will never take over my developer job! AI.... just coded a more performant and lightweight adobe suite in rust? darn...

Unlimited slop works! Title kinda misleading (at least appears so to me). No rev. eng. (so "cleanroom" to be relevant) takes place. "ACS clone" would be more appropriate.

Reply on news.ycombinator.com