Brandon Thomas says his AI-built Adobe clones are not close to ready
InfoQ reports that Brandon Thomas, the developer behind seven AI-built open-source Adobe clones, told Hacker News the apps are not anywhere close to ready, and the project's own roadmap puts readiness for real work at about 45 percent.
Image: InfoQ
Why it mattersA team deciding whether to try an AI-built replacement for a working tool should read the maker's own numbers before switching, because the roadmap in this project is harder on the code than most of its reviewers were.
A developer who published open-source replacements for seven Adobe applications told Hacker News they are "not anywhere close to read" and asked people to wait for the beta. InfoQ reported this on 22 October in a piece by Steef-Jan Wiggers, which quotes the Hacker News comment alongside the project's own roadmap. Brandon Thomas had posted earlier across five subreddits with a different message: beta and roughly 100 percent parity within a month. The repository sides with the Hacker News version.
Thomas used Claude Opus 5.5 to build the applications in Rust, with WebAssembly builds for the browser. The seven are PhotoCraft, VectorCraft, EffectCraft, LightCraft, PDFCraft and alternatives to Premiere and InDesign, mirroring Photoshop, Illustrator, After Effects, Lightroom, Acrobat Pro, Premiere and InDesign. All of it ships under MIT and Apache licences.
What the roadmap says
InfoQ cites the ROADMAP.md file, updated 11 October. Feature breadth sits at about 76 percent of the target application, from four counts: 628 of 628 menus, 56 of 68 tools, 30 of 35 panels, and 13 plus 4 of 31 formats. Readiness for real work is a separate number, and it is lower. The roadmap puts it at about 45 percent. A mainstream practitioner gets 43 percent and someone using only core features gets 61 percent. Beta is 1,100 to 1,800 agent-hours out and full parity 2,300 to 4,100, counted as Claude Opus 5.5 wall-clock hours for one agent working in sequence. The target is Photoshop 2026 version 27.11.0, measured from the installed application on 10 October.
Users of versions 0.5 and 0.6 have filed 579 open issues, according to InfoQ. Twenty-two of twenty-five performance budgets are missed. And the roadmap says files PhotoCraft writes "are not yet reliably accepted by Photoshop". For software meant to replace Photoshop, that is not a rough edge.
The 317 contributors
InfoQ reports the git history. First commit 30 September, latest 13 October. In those fourteen days the repository took 2,152 commits, 1,004 merged pull requests and 539,017 lines of Rust across 1,306 files. Git records 317 distinct contributor identities, so the "solo developer" framing stopped being true almost at once. Claude appears as a named commit author 277 times.
InfoQ records how people who installed the apps split along the roadmap's line. One Reddit commenter said the Photoshop clone cannot even eye-dropper with the Alt key. Photographers hit the RAW pipeline, with noise and banding on a Sony a7c file and no RAW support for several Nikon bodies. The simplest app did best: on Hacker News, one commenter reported PDFCraft working without bugs.
InfoQ also quotes the developer of Hew, an open-source SketchUp replacement built with Claude, who said the first version came out fast and looked great but failed in unexpected ways, and that each major feature now takes several hours of Claude work plus many more hours of their own testing. They put parity at a year. Thomas has adjusted his own claim too, telling Hacker News that "99% parity will take a while, but I'm sure it'll be measured in months and not years."
For a reader thinking about any AI-built replacement for a tool they use every day, the readiness number is the one worth reading. Counting menus and tools reaches 76 percent in two weeks. Making the files the clone writes work in the application it replaces takes much longer.
Source
This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.
Get AI News in your inbox
New developer tools, model and agent releases, and how teams are actually using them to release software. Short, and only when there is something worth reading.
