Adobe in Trouble After a Developer Reverse-Engineered Its Creative Suite and Released It for Free to Everyone
Adobe’s newest headache isn’t just another startup nipping at its heels, it’s a developer in Atlanta who rebuilt seven of Adobe’s core apps, made them open source, and gave them away for free after getting dinged with a cancellation fee. That’s the kind of petty, brilliant, very-online move that turns a subscription gripe into a company problem. And because he used AI to do it fast, the real story isn’t just whether Adobe can sue, it’s whether Adobe’s whole “rent forever and hope nobody gets off the treadmill” model is starting to crack.
Brandon Thomas, an engineer and filmmaker based in Atlanta, says the trigger was Adobe’s cancellation fee, which hit him more than once. Not a pricing annoyance, not a feature complaint, the bill for trying to leave. That’s the detail that matters, because it explains why this wasn’t just a hobby project or a flex. It was a revenge project with source code.
And Thomas didn’t make one cute clone and call it a day. He pushed out PhotoCraft, VectorCraft, FilmCraft, LightCraft, EffectCraft, DesignCraft, and PrintCraft, which he later renamed PDFCraft. In plain English, that’s Photoshop, Illustrator, Premiere, Lightroom, After Effects, InDesign, and Acrobat. Seven apps. In about a week. If that sounds impossible, that’s because under normal circumstances it is.
The part that makes this story weirdly dangerous
Thomas isn’t a random guy who installed Python last Tuesday and woke up feeling blessed. He says he’s been in software for around 15 years and has spent about 10 of those years writing Rust, which is a programming language people use when they want speed, memory safety, and fewer catastrophic bugs. He also founded FakeYou, the AI celebrity-voice app, and later an AI film startup that became Artcraft. So when he says software can be rebuilt faster than the old guard expects, he’s not guessing. He’s been playing with automated creativity for years.
The timeline alone is ridiculous. He says the first line of code landed on September 30. The first release of PhotoCraft arrived on October 2. By October 7, he had posted the project on Reddit and Hacker News, which is basically how you invite the entire internet to climb over your fence and kick your tires. And kick them it did.
Across the public code history for all seven apps, there are around 7,500 changes in roughly 10 days, and about 60% of them list Claude as co-author. Claude is Anthropic’s AI model, and Thomas used it to help generate and shape the code. That matters because it changes the frame from “one smart dev shipped a miracle” to “one smart dev orchestrated a machine-assisted sprint that would’ve taken a traditional team much longer.”
By the standards of software, that’s not just fast. It’s destabilizing. A lot of legacy companies are built on the assumption that cloning their product would take too many people too much time. This story says, maybe not anymore.
What PhotoCraft and friends actually do
The seven apps are clearly meant to map onto Adobe’s creative stack, but they aren’t cardboard props. They run on Mac, Windows, and Linux, and they even work in a web browser. That alone is a big deal, because Adobe users have spent years living with platform lock-in, the annoying little reality where your toolset follows Adobe’s rules, not yours.
They’re open source under the MIT and Apache licenses. Those are permissive licenses, which means anyone can download the code, use it, modify it, and redistribute it. Unlike proprietary software, there’s no paywall around the source itself. And unlike “free trial” software, there’s no ticking clock threatening to lock you out later.
Each app points at a different part of Adobe’s empire:
- PhotoCraft for Photoshop-style image editing
- VectorCraft for Illustrator-style vector graphics
- FilmCraft for Premiere-style video editing
- LightCraft for Lightroom-style photo organization and editing
- EffectCraft for After Effects-style motion and compositing work
- DesignCraft for InDesign-style layout and publishing
- PrintCraft, now renamed PDFCraft, for Acrobat-style PDF work
That list is the whole game. Adobe’s power isn’t one app, it’s the ecosystem. If you can recreate the entry points, even roughly, you can start attacking the lock-in from multiple directions at once. That’s why this project got attention so fast. Nobody was waiting for another “Photoshop but blue” demo. They were looking at a seven-part challenge to the entire bundle.
Is it actually good, or just good enough to scare Adobe?
The honest answer is somewhere between “impressive” and “not remotely finished.” One reviewer tested PhotoCraft on a weak laptop and found it fast. It supports layers, masks, smart objects, brushes, and it opens Photoshop files. Those aren’t toy features. They’re the basic language of real image editing.
But it’s not Photoshop yet, and nobody serious is pretending otherwise. Blend modes and filters don’t match Adobe’s behavior, and the reviewer saw it eat about 5 gigabytes of memory while opening a single file. That is both absurd and, annoyingly, very on-brand for Photoshop-like software. If you’ve ever watched a creative app turn your machine into a space heater, you know the vibe.
The verdict from that testing was blunt: not for professionals. That doesn’t mean useless. It means the gap between “works” and “production-safe” is still huge. Professionals care about predictable output, weird edge cases, plugin support, file fidelity, and all the invisible junk that only becomes visible at 3 a.m. the night before a deadline.
Thomas has been clear that this is a super early alpha. He originally talked like Adobe parity would arrive within a month, then walked that back to months, not years. That’s a much more realistic promise, and honestly, a more interesting one. The point isn’t that the apps are done. The point is how quickly they got to “usable enough to matter.”
The business model behind the free apps is not as pure as it sounds
There’s a small but important catch here. The apps themselves cost nothing, but the company behind them is still trying to make money. The plan is to sell AI image and video credits, plus cloud storage on top. So this isn’t a charity project, and it’s not some open-source monastery where nobody has to pay rent.
That matters because “free” in software usually means one of three things: the code is free, the product is free, or somebody else is paying and hasn’t noticed yet. Here, it’s the first two, with a business model hanging off the side. That’s a very modern arrangement, and it’s probably the only way projects like this survive long enough to become annoying to incumbents.
Thomas also isn’t exactly a lone hacker in a basement. He’s coming out of a funded startup, and he has a small team plus dozens of volunteers now contributing code. That doesn’t make the achievement smaller, it just makes it more honest. We should stop pretending these things are always one-person myths. A lot of the time they’re small teams using AI as a force multiplier, and that changes the tempo of software competition.
He’s also already talking about bigger targets. PowerPoint, Excel, AutoCAD, and Pro Tools are on the list. That’s not a casual aside. That’s a declaration that the same method could be pointed at office software, CAD tools, and audio workstations too. Adobe isn’t the end goal. It’s the opening act.
Why Adobe looks rich and terrified at the same time
Adobe’s numbers, on paper, still look enormous. In its last quarter, it reported record revenue of $6.7 billion. It says more than a billion people use its products every month. That’s not a dying company. That’s a company with a giant installed base and a huge amount of inertia.
But the stock market has clearly stopped believing the old story. In November 2021, Adobe shares were around $680. Last week, they were around $240. That’s roughly 65% gone. Goldman Sachs even put a rare sell rating on the company, which is the sort of thing that makes executives sit up straighter in meetings.
The other warning sign is growth. Adobe’s new subscription business has slowed, falling by about a third. And after 19 years in charge, the company’s longtime boss is stepping down in December. When the cash machine is still humming but the market starts treating the whole setup like a mature, fragile utility, you know people are staring at the same cliff from different angles.
That’s the real tension here. Adobe is still huge. It’s also increasingly hard to argue that it’s unstoppable.
Adobe spent years making itself easy to hate
This is the part that gives the whole story teeth. Adobe didn’t just get unlucky. It built a customer base and then spent years teaching that customer base to resent the company.
The pivot to subscriptions is the obvious starting point. In 2013, Adobe stopped selling its software outright. Before that, the Master Collection was a one-time purchase, about $2,500, and then it was yours. After the switch, it became a rent relationship. The base plan started at $50 a month. It’s now $70 a month, or $105 month-to-month. If you’re a freelancer or a small studio, that’s not pocket change. It’s a recurring overhead item that never goes away.
And it wasn’t just the price. Tens of thousands of people signed a petition against the move and Adobe did it anyway. That kind of decision creates a memory in the market. People may keep paying, but they stop trusting your future promises.
Then there was the 2013 breach, where hackers stole details on 38 million Adobe users and the source code for Photoshop. That’s a brutal combo. User data plus product code is the kind of thing that raises questions about everything from account security to software integrity. When a company’s security story gets linked to one of its crown-jewel products, it’s hard to cleanly separate the brand from the mess.
In 2019, Adobe pulled older versions of its apps and warned users that they might face potential claims of infringement for continuing to use software they had already paid for. That’s the sort of thing that makes normal people stare at the screen and ask what, exactly, ownership is supposed to mean. Imagine buying a car and later being told the manufacturer doesn’t like the model year you’re driving.
That same year, Adobe shut down every account in Venezuela because of U.S. sanctions. Refunds only came after backlash. Again, the specific policy is less important than the pattern: users learn that access can disappear, plans can change, and the company will always be on the side of the contract, not the customer.
Then came 2022, when Pantone colors in people’s files turned black unless they paid for another subscription. That one hit designers right in the gut. Pantone isn’t some random feature. It’s a color system used in print and brand work, the thing that keeps a logo from looking right on one run and wrong on the next. Turning those colors black unless users paid again was the kind of move people remember for years.
And finally, the cancellation fee itself. If you signed up for the annual plan, paid monthly, and tried to leave early, Adobe could charge you half of what remained on the year. That’s the policy that pushed Thomas over the edge. The U.S. government sued Adobe over it in 2024, alleging the fee was buried in fine print and that canceling had been made deliberately hard. The suit also named two executives personally. This March, Adobe agreed to a $150 million settlement, split into $75 million in penalties and $75 million in free services. Adobe says it did nothing wrong, and the UK competition watchdog has opened its own investigation.
That’s not a random pile of bad luck. That’s a company that trained people to expect friction, then acted surprised when the friction started to define the brand.
Adobe’s AI story didn’t exactly help
In 2024, new terms popped up saying Adobe might access user content through automated and manual methods. Creators freaked out, because that sounds a lot like “we may look at your files,” which is not the vibe you want from a company that sells you creative software. Adobe later rewrote the terms and promised it wouldn’t train on user work. Fine. But the trust damage was already doing pushups.
At the same time, Adobe has been pushing Firefly as its ethical AI, the one trained on licensed or company-owned material. That pitch would’ve landed better if Bloomberg hadn’t reported that some training images were made by Midjourney. Stock contributors also apparently couldn’t opt out in the way they wanted, and the estate of Ansel Adams publicly told Adobe it was officially on their last nerve. When a photographer’s estate is irritated enough to say that in public, you’re not in a good place.
This matters because Adobe wanted to be seen as the safe AI company for creatives. Instead, it’s become another company asking artists to trust the machine while the machine company explains the fine print. That’s not a stable moral position, especially in a market where alternatives are getting decent fast.
The competition is no longer embarrassing
Adobe used to benefit from the fact that the alternatives were clunky, narrow, or weirdly niche. That’s changed.
Canva has more than 260 million users. Affinity, which many designers used as the “serious but cheaper” answer to Adobe, went completely free last year. DaVinci Resolve has had a free version for years and is good enough that plenty of editors happily live there. Full disclosure, the video this topic came from was edited in Resolve, not Premiere. That’s not a minor detail. It’s the sort of quiet adoption that happens when tools stop being symbolic and start being practical.
Adobe itself seems to know the pricing wall has hit the ceiling. On its last earnings call, the boss said he was glad they had deferred pricing changes in favor of user growth. Corporate translation: we can’t keep ratcheting the price forever and pretending nobody notices.
If you want the cleanest expression of the pressure Adobe is under, it’s this: free alternatives are no longer jokes. They’re good enough for real work, good enough to start a migration, and good enough to make subscription fatigue actionable.
Can Adobe sue, and would it even matter?
This is where the legal nerd stuff gets interesting. On copyright, Adobe’s path is shakier than a lot of people assume. In 1995, a U.S. court said a software menu structure is a method of operation, which means it can’t be copyrighted the way a novel or song can. Apple once tried to own the look and feel of a desktop and lost. In 2021, the Supreme Court sided with Google against Oracle over copying how software talks to other software. And courts have also ruled that taking apart a product to build a compatible one can be fair use.
That last point is huge. Adobe publishes a specification for the Photoshop file format, and Thomas says this is a clean-room rebuild, with no Adobe code. Clean room here means the software is recreated from public behavior and specifications, not copied from proprietary source code. If that’s true, the copyright attack gets harder.
That doesn’t make Adobe powerless. Reviewers have already said some of these apps look almost identical to Adobe’s, which opens the door to trade dress arguments, meaning claims about copying the look and presentation of a product. Adobe also owns thousands of patents. Patents are a different beast from copyright, and they can be used to attack specific implementation details rather than the whole product. If Adobe wants a fight, it has tools.
But the legal terrain isn’t friendly in the way a lot of executives would like. And the optics are brutal. A company that just settled a $150 million cancellation-fee case probably doesn’t want to spend the next year explaining why it sued a developer for making a free alternative after getting fed up with that same fee.
Why open source copies are harder to kill than people think
There are already free Photoshop alternatives. GIMP has been around since 1996. Photopea is another one, built by a single developer and running in the browser. Adobe has apparently never taken either to court, which tells you something. Sometimes the most effective response to a rival is not litigation. It’s pretending they’re beneath you and hoping the market agrees.
That strategy only works until the rival becomes useful enough to matter. Once that happens, the distribution matters. PhotoCraft has already been copied nearly 6,000 times in a few days. With open source, you can’t claw those copies back just because you found the project annoying. The MIT and Apache licenses don’t work that way. Once the code is out, it’s out.
That’s why the comparison people keep making to Nintendo, YouTube-dl, and old DVD-cracking cases only goes so far. Nintendo went after the Switch emulator Yuzu and got a $2.4 million settlement, plus more than 8,500 copies removed from GitHub. But that was about encryption circumvention and piracy. This is a from-scratch creative tool. Different facts, different problem.
YouTube-dl got pulled from GitHub and came back within a month. Years ago, when Hollywood tried to ban code that cracked DVDs, people printed it on T-shirts. The lesson isn’t that the law never matters. It’s that code has a nasty habit of spreading faster than legal threats can contain it once enough people decide it’s useful or symbolic.
What happens next for Adobe and for everyone else
The likely short-term outcome is boring in the way big-company crises often are. Adobe probably ignores this first, the way it ignored GIMP and other challengers. If it does respond, a lawyer letter asking Thomas to tweak the icons, layout, or branding is more likely than some dramatic courtroom showdown. That would be annoying, but it wouldn’t stop the code from existing.
A full patent war is the spicy scenario, but it’s also the one most likely to backfire. If Adobe starts swinging hard at a free alpha that exists because its cancellation policy annoyed somebody, it risks turning a niche project into a cause. That’s not a lawsuit. That’s a marketing campaign for the other guy.
The more sensible move, and the one Adobe is already drifting toward, is to compete on terms people actually like. Cheaper plans. Easier cancellation. A free tier that doesn’t feel like a trap. Better trust around AI. Less of the “you can leave, but you’ll be sorry” energy. The company’s free-user count reportedly jumped 70% last year, which suggests it understands the future is not just about squeezing existing customers harder.
And that’s the punchline. Adobe can probably survive as a company. Big studios, agencies, and teams don’t switch creative pipelines overnight. They have plugins, habits, training, file archives, and years of muscle memory. But Adobe may not survive as the kind of company that assumes captivity is the same thing as loyalty.
Thomas’s project is rough, early, and not ready to replace the real thing for professionals. None of that matters as much as the fact that it exists at all. If a small team with AI help can rebuild the shape of seven Adobe apps in about 10 days, the barrier to entry just changed. The next person won’t need Adobe’s permission, and they probably won’t need as much time either.
That’s the part Adobe should be worried about. Not one clone. The method. The recipe. The fact that the cat isn’t just out of the bag anymore, it’s teaching other cats how to open bags.
So no, this isn’t the end of Adobe. But it might be the end of Adobe thinking it can charge forever, trap people with fees, and expect gratitude for the privilege. That era looks shaky now, and honestly, it should.