Brandon Thomas’s ArtCraft project has grown from a set of seven open-source alternatives to Adobe creative apps into a second effort targeting Microsoft Word, Excel and PowerPoint. The expansion shows how AI coding tools can help turn subscription frustration into competing software—but the new apps remain early-stage, and their reported feature coverage is not independent proof that they are ready to replace established products.
2
4
7
From Adobe alternatives to Office apps
Reporting links the project’s origins to Thomas’s frustration with Adobe’s subscription and cancellation practices.
35 In early October, ArtCraft released Rust-based alternatives modeled on seven Adobe applications: Photoshop, Illustrator, Premiere Pro, Lightroom, Acrobat, After Effects and InDesign. The project says the tools were developed with Anthropic’s Claude Opus 5.5.
2
8
The project then broadened its scope with three open-source applications: WordCraft for word processing, GridCraft for spreadsheets and DeckCraft for presentations. These target Word, Excel and PowerPoint, respectively, and reporting says they were also built with Claude.
1
7 The names and similarities identify the products they aim to replace; they do not make them Microsoft applications or establish feature equivalence.
What the reported feature numbers do—and don’t—show
Some coverage cites substantial feature counts: GridCraft is reported to support 513 of Excel’s 523 worksheet functions, while DeckCraft is said to cover about 79% of PowerPoint’s features.
20
23 Those figures are reported claims, not independent validation of accuracy, compatibility or reliability across real-world work.
That distinction matters for office software. A function count alone cannot show whether spreadsheets calculate correctly in varied cases, or whether documents and presentations preserve their formatting when exchanged with other applications. One repository describes WordCraft as reading and writing DOCX files and running on several desktop platforms and in a browser; that is a claim about WordCraft, not proof of equivalent support across all three apps.
21
Coverage also describes the Office applications as alpha or early-stage software, not finished replacements.
4
27 ArtCraft has shared an optimistic view of how the tools worked internally, but that assessment comes from the project rather than independent testing.
34
Subscription frustration as a software-development motive
ArtCraft is an example of a complaint about a commercial product becoming a public software project. Engadget connected Thomas’s initial effort to resentment of Adobe’s subscription model, and The Week described this broader kind of response as “open-source retaliation.”
34
35
AI-assisted coding may make it easier to produce recognizable alternatives, but a fast first release does not establish that an application can match the depth, dependability or workflow fit of a mature product. For now, ArtCraft is evidence that dissatisfaction can motivate attempts to compete—not that subscription software has already been displaced.
Legal questions remain separate from the project’s claims
ArtCraft describes its approach as a clean-room reimplementation, but that description is not a legal ruling or independent verification of how every part of the software was made.
39 Legal analysis of the Adobe alternatives distinguishes reverse engineering, which may generally be lawful in the United States, from separate questions such as license terms, patents and copying a product’s name or appearance. The specifics matter.
48
There is no public court ruling establishing that ArtCraft infringes Adobe’s rights, according to one legal-risk analysis.
55 At this stage, warnings about a possible lawsuit are predictions, not findings that the project is unlawful. Nor does an open-source license, by itself, resolve every intellectual-property question.
The practical takeaway
ArtCraft’s expansion is significant as a demonstration of AI-assisted software development and user-led alternatives to subscription products. But the Adobe and Office tools should be judged as separate, evolving applications—not assumed to be full substitutes because they share familiar names, interfaces or feature counts. The current evidence supports curiosity about the project, alongside caution about its maturity and unresolved legal questions.
2
4
48