Brandon Thomas Rewrites Seven Adobe Apps with Claude Opus 5.5; REA Project Nears 100k Stars
Key Highlights
Frustrated by Adobe's subscription cancellation penalties, developer Brandon Thomas used Claude Opus 5.5 to reverse-engineer and rewrite seven Adobe applications, including Photoshop, in Rust, releasing them all open-source and free. The Photoshop project alone gathered 46,000 stars in under ten days, and the broader REA (Rust-rewritten Adobe) project is approaching 100,000 stars. What makes this explode is that it turns an individual using AI to rewrite commercial software from a joke into reality.
What It Does
The REA project rewrites the core functions of several flagship Adobe applications in Rust. The goal is not pixel-perfect cloning but providing open-source, free, locally-running alternatives to Adobe's much-loved-yet-resented subscription tools. For users fed up with not being able to open their own files without renewing, this is an emotionally resonant substitute. AI here acts as an accelerator, compressing what once required a team years of work into results one developer can ship in months.
Technical Details
The choice of Rust matters: it is memory-safe and performant, suited to desktop software like image processing that demands stability and speed, and it compiles to cross-platform native binaries without the performance constraints of a web version. Using Claude Opus 5.5 to assist the rewrite shows that top coding models can now take on the complex task of understanding large closed-source software behavior and producing runnable alternatives. The 46k stars in under ten days confirms the open-source community's intense desire to escape subscriptions.
Versus Alternatives
Versus the commercial Adobe suite, REA's advantages are free, local, and unshackled from subscriptions; the trade-off is that functional completeness and ecosystem (plugins, format compatibility) still trail software polished for decades. Versus other open-source design tools like GIMP or Krita, REA's differentiator is using AI to quickly cover Adobe's usage habits, lowering migration cost. On the legal front, reverse-engineering behavior rather than directly decompiling source is its delicate compliance footing.
Industry Impact and Use Cases
REA is a signal: as top coding AI proliferates, the barrier for individual developers to challenge commercial-software giants is collapsing. The lock-in effect of subscription software may be weakened by open-source-plus-AI alternatives. For Adobe and peers, it is a reminder that once users' dissatisfaction is cashed out in technology, moats are not necessarily reliable. For ordinary users, it demonstrates the possibility of using AI to build your own tools and adds fuel to the debate over whether software should be this expensive and this binding.
The broader lesson of REA is that the economics of software labor are shifting underneath the subscription model. When a single determined developer, armed with a frontier coding model, can produce credible alternatives to products built by hundreds of engineers, the historical justification for steep recurring fees weakens. This does not mean commercial software disappears; it means the bargaining power tilts, and incumbents must justify their prices with capability and ecosystem rather than lock-in. For the open-source movement, REA is a morale boost and a proof point. For Adobe specifically, it is a signal to take community frustration seriously before it crystallizes into durable, popular alternatives that erode the long tail of users who are only loosely committed.
Still, the legal and ethical lines deserve care. Reverse-engineering behavior to build compatible, independently written implementations sits in a different place than copying proprietary code, but the boundary is contested and jurisdiction-dependent, and the REA project will be watched closely by lawyers as much as by developers. The community's enthusiasm should be paired with clarity about what is and is not permissible, so that the project's momentum is not derailed by a single misstep. If it navigates that carefully, REA could become a template for how determined individuals, augmented by coding AI, contest entrenched software monopolies, and a case study in the changing economics of creative and productivity software.
For users tempted to switch, the pragmatic advice is to test REA on the specific Adobe tasks you actually do, not on the demos that look impressive, because coverage is uneven and will be for a long time. Where it works, it is a genuine escape from subscription fees; where it does not, the commercial tool remains the safer bet. The project is best understood as a catalyst and a proof of possibility rather than a finished replacement, and approaching it that way avoids disappointment while still capturing the parts that are already excellent.
The cultural ripple may outlast the code. REA has given permission, tacitly, for ordinary developers to imagine replacing the tools they resent, and that shift in expectation is hard to unsee once it spreads. When users start from the assumption that an open alternative could exist, vendors lose the comfort of captive demand and must compete on merit. Whether REA itself becomes the replacement or merely the spark, it has already changed the conversation around software ownership, and that change will keep echoing through how the next generation of tools gets built and priced.