A startup called Stacklet built an iOS app that let users create functioning apps by describing them in natural language. Apple rejected it from the App Store, citing guideline 4.7 which prohibits apps that “create standalone apps or experiences.” Instead of giving up, the team rebuilt the entire product as an iMessage extension. Apple cannot block that.
What Stacklet Originally Did
The concept was simple: you describe what you want in plain English (“make me a workout tracker with a calendar view and progress charts”), and Stacklet generates a working mini-app on your phone. No coding knowledge required. The generated apps ran inside Stacklet’s container, used local storage, and could be shared with other Stacklet users.
This is what developers call “vibe coding,” a term that emerged in 2025 for the practice of building software through natural language descriptions rather than writing code. Tools like Cursor, Replit, and Bolt popularized the approach on desktop. Stacklet brought it to mobile.
Apple’s rejection was predictable but still frustrating. The App Store guidelines are designed partly to maintain Apple’s control over what software runs on iPhones. An app that generates other apps threatens that control, even if the generated apps are sandboxed and harmless.
The iMessage Pivot
iMessage extensions operate under different rules than standalone apps. They can display interactive content, process user input, and maintain state across conversations. Stacklet’s team realized they could deliver most of their product’s functionality through an iMessage extension that generates interactive “cards” instead of standalone apps.
You text Stacklet’s iMessage bot: “budget tracker for my trip to Japan.” It generates an interactive card with expense categories, currency conversion, a running total, and a simple chart. The card lives in your iMessage conversation, shareable with travel companions, and functional without leaving the Messages app.
The technical limitations are real. iMessage extensions cannot access the camera, GPS, or notifications. But for simple utility tools, the format works well enough that users do not notice the constraints. If you have been following how AI tools are finding creative distribution channels, this iMessage approach is clever.
Apple’s App Store Problem
Stacklet’s story highlights a growing tension. Apple wants iPhone users to have access to AI tools. Apple also wants to control what software runs on iPhone. These goals increasingly conflict as AI makes it possible for anyone to create custom software on the fly.
Other vibe-coding tools have hit similar walls. Replit‘s mobile app had features restricted. Several code-execution apps have been rejected or required modifications. Apple’s 30% commission on in-app purchases adds a financial dimension: if users generate their own mini-apps, they bypass the App Store revenue model entirely.
The iOS 26 update introduced some sideloading provisions in the EU, which could eventually provide an alternative distribution path. But for now, in most markets, the App Store is the only way onto iPhones.
Should You Try It
If you have an iPhone and use iMessage, Stacklet’s extension is free to try for basic tool generation. Premium features (more complex tools, shared collaboration, persistent storage) cost $4.99/month. For quick utility tools like messaging productivity features and calculators, it is surprisingly capable.
The broader takeaway is that AI-powered creation tools will keep finding workarounds when platforms block them. The demand exists. The technology works. Distribution is the only bottleneck, and entrepreneurs are resourceful about solving distribution problems.
