Recommended App Store distribution strategy for apps that require Foundation Models

Hello, @a_okano! The recommendation on the App Store side is to provide some baseline functionality to all users, regardless of whether Apple Intelligence is available. The App Store doesn’t support a required device capability for Apple Intelligence. Even on compatible devices, there are a number of reasons why Apple Intelligence could be unavailable, such as if the user selected an unsupported Siri language, is located in an unsupported region, or opted out of Apple Intelligence.

Thank you for the clarification.

One follow-up question:

For an app whose primary purpose is an AI agent, could you clarify what Apple considers to be an appropriate "baseline functionality"?

For example, would a limited non-AI experience generally be expected, or is it acceptable to explain that the primary feature requires Apple Intelligence?

I'm trying to better understand the recommended UX for this scenario.

@Frameworks Engineer

Nearly every outside developer will tell you it’s not acceptable to coerce developers to deliver subpar experiences to users.

What, the user then has to go through the refund process? No one wants that.

The App Store should support mechanisms so users don’t accidentally download apps they can’t get the most out of.

Don’t worry. I know you’ll take your 30% anyway.

Apple’s approach makes sense because Foundation Models may not be available for every user, even on a supported device.

The app should use Apple’s on-device model when it is available, but it should still offer some useful functionality when it is not. For an AI app, that could mean falling back to a cloud model. This avoids users downloading an app only to find that its main feature does not work on their device.

Hopefully I can help clarify!

As of WWDC 2026, Foundation Models framework covers both on-device foundation models and server-based models.... and both Apple Foundation Models as well as any other LLMs. So "foundation models" can mean a bunch of different things and a bunch of possible models, which is part of the reason why there isn't currently a clean device-capability flag.

The full list of Apple Intelligence requirements (for Apple Foundation Models) can be found here https://support.apple.com/en-us/121115 and include a combination of regional and hardware requirements.

Models from other sources can be used with Foundation Models using MLX or CoreAI, so you can still reach users with hardware that can't run Apple's on-device foundation model.

So... what can you do?

1. Run an availability check as soon as you launch your app. For example https://developer.apple.com/documentation/foundationmodels/systemlanguagemodel/isavailable for the on-device model. Availability can tell you additional information about compatibility... like if the model is downloading or not available for that language. From a UX standpoint, try to check availability before anyone agrees to pay for your app's service, to avoid someone paying for what they can't use.

2. Figure out if you can use a different model as backup, if Apple's on-device foundation model isn't compatible with the device. Any server or local LLM might do.

Let me know if there's any point I can help clarify or discuss more, thanks!