Developers allege Apple blocked months of iOS updates to Replit and vibe‑coding tools

Here’s the cleanest way to understand the dispute. 1/ Developers behind AI “vibe-coding” apps say Apple spent months blocking iOS updates to tools including Replit and Vibecode while citing existing App Store rules against downloading or executing code that changes an app’s functionality. Apple has said it has no rule specifically targeting “vibe coding” apps, and pointed instead to Guideline 2.5.2. (macrumors.com) 2/ The core issue is not “AI” in the abstract. It is how these apps work on iPhone. Apps like Replit let users generate software from prompts. Apple’s review teams objected when the generated app was previewed inside the original iOS app, typically through an embedded web view, according to reports citing people familiar with the matter. Apple’s position was that this could amount to code changing app behavior after review. (macrumors.com) 3/ Replit became the highest-profile example. MacRumors, citing The Information, reported on March 18 that Apple had blocked updates for Replit and Vibecode unless they made changes. Replit had argued it had been on the platform for years under the same rules. (macrumors.com) 4/ Apple’s written rationale was old, not new. The company told reporters in March that the restriction came from longstanding App Store rules, especially 2.5.2, which says apps must be self-contained and may not download, install or execute code that changes features or functionality. Apple also said the guidelines were meant to preserve user safety while allowing innovation. (macrumors.com) 5/ The workaround developers described was also pretty specific. Reports said Apple was more likely to approve updates if generated apps opened in an external browser instead of being previewed inside the iOS app itself. In other words: move the live preview out of the native app shell and into the web. (macrumors.com) 6/ That helps explain the “months blocked” claim now circulating on X. Replit’s iPhone app got its first update in four months on May 15 after what 9to5Mac and The Verge both described as an App Store review dispute. The Verge said Replit CEO Amjad Masad posted that the company had “worked things out with Apple.” (9to5mac.com) 7/ The timing matters. If the update shipped on May 15 after four months without one, that places the freeze back to roughly January. Separate March reporting said Apple’s pushback had already become visible by March 18, when Replit and Vibecode were publicly identified as affected. That does not prove every app was blocked for the same duration, but it does support the broader complaint that developers were dealing with a prolonged review bottleneck. (9to5mac.com) 8/ Replit says the delay held back product releases on iPhone. Coverage of the May 15 update said features such as Agent 4 and other collaboration or project-viewing improvements had been announced earlier but only reached iOS after the review issue was resolved. (9to5mac.com) 9/ This was not limited to Replit. March coverage said Vibecode was also blocked from releasing updates unless it changed features, and later reporting said Apple had also removed the vibe-coding app Anything from the App Store. That suggests a category-wide enforcement push rather than a one-off dispute with a single developer. (macrumors.com) 10/ The open question is consistency. Developers’ complaint is not just that Apple enforced 2.5.2. It is that Apple appears to be applying an old rule to a new class of AI products in ways they say are uneven and hard to predict. Apple’s public answer so far is narrower: there is no anti-vibe-coding rule, only existing review guidelines. (macrumors.com) 11/ Where things stand now: Replit appears at least partially unblocked, with a fresh iPhone update live as of May 15. But the broader policy question — whether Apple will keep forcing these tools to push previews and execution flows into the browser — still appears unresolved. (9to5mac.com) 12/ So the story is less “Apple banned AI coding” than “Apple used old App Store code-execution rules to constrain how AI coding apps can work on iPhone.” That distinction is why developers are calling it gatekeeping and Apple is calling it standard review enforcement. (macrumors.com)