Apple blocks Replit and Vibecode updates in the App Store over vibe-coding features

Apple has temporarily stopped publishing updates for several apps tied to vibe coding, including Replit and Vibecode. Based on the available description of the situation, the objections concern specific features that the company considers incompatible with App Store rules.

What happened

What got blocked were new versions of the apps, not the idea of mobile AI-assisted development itself. That distinction matters: Apple has not declared war on the whole segment, but it has shown that it is closely examining how such products work under the hood.

For services that promise to generate, run and refine code with almost no user involvement, iOS has long been a difficult platform, because of strict rules on code execution and on changing an app's behavior after review.

Replit and Vibecode became prominent examples because both services sell the idea of development "from your phone, from a description." The user states a task in natural language, gets a project scaffold, and can then edit the interface, logic and file structure. For a mass audience, that is what vibe coding means: less manual engineering, more dialogue with a model and fast iteration.

The story also matters because mobile clients have become more than a web add-on for these platforms: they are part of the core funnel. The iPhone makes it easier to show the product's magic: open the app, describe an idea, get a prototype. If exactly this scenario starts to clash with store rules, companies have to rebuild not just the interface but the whole logic of onboarding, monetization and user retention.

Where the conflict arises

Apple's exact wording of its objections has not been disclosed in the publication, but the logic of the conflict is clear. The App Store has historically been wary of scenarios where an app turns into a shell for an external execution environment, or can radically change its functions after the review team has checked it. When an AI tool doesn't just write text but assembles an interface, runs code, connects third-party services and shows the result inside the mobile client, it moves into territory where Apple's rules are especially strict.

The problem here is probably not code generation itself but the degree of autonomy. A chat assistant that suggests a snippet of a React component is one thing. A product that builds a working app from a single instruction, lets you run it, and changes the service's behavior almost in real time is another. For the App Store, that is no longer just an editor but potentially a new platform inside a platform, and Apple traditionally tries to keep such scenarios under full control.

For such products, the high-risk zone usually looks like this:

• generating and running user code directly from the mobile client
• remote changes to app behavior without a separate review of each new feature
• access to external interpreters, builders or containers hidden from moderation
• scenarios where AI creates full mini-apps inside an already published shell

What happens next

The most likely scenario is that developers will cut the disputed capabilities or move them out of the iOS app. For example, they might keep a client in the App Store for chats, project viewing and basic editing, and move full generation, deployment or execution to the web version. This is a typical compromise on Apple platforms: the mobile app becomes a storefront and remote control, while the most sensitive functions live in the browser, where App Store restrictions do not apply.

For users, this means less a loss of the service than a slower pace. New features may ship first in the browser, on desktop or via TestFlight, and reach the App Store later and in trimmed-down form. For the companies themselves the risk is higher: the vibe coding market is growing precisely on the promise of "build an app in a couple of minutes," and any platform restriction breaks that effect of simplicity on which the growth funnel is built.

Startups that build their product around mobile UX and a quick wow effect take a separate hit. If critical scenarios cannot be legally packaged into an iOS app, they will have to explain to users why the "real magic" only begins after switching to the browser. This raises friction, lowers conversion and widens the gap between how the product looks in a promo video and how it works on a real device.

What this means

Apple has shown that it is ready to rein in even popular AI products if they stray too close to scenarios prohibited on iOS. For the market, this is a signal: the future of vibe coding depends not only on model quality but also on how carefully services package code generation, execution and publishing within the rules of the major mobile platforms.