What devices support the Foundation Models framework?

It requires Apple Intelligence-enabled hardware such as iPhone 15 Pro or newer, and M1+ iPads and Macs running iOS/macOS 26 or later.

Can I use this for general knowledge queries like ChatGPT?

No, the on-device 3B model is specifically optimized for tasks like summarization and classification rather than broad world knowledge.

Does this require an internet connection?

No, one of the primary benefits is that all inference is performed entirely on-device, ensuring privacy and offline functionality.

How does this skill handle context limits?

The skill provides guidance on checking input length and proactively trimming content to avoid context overflow and silent failures.

What is the benefit of using @Generable?

The @Generable macro provides type-safe structured output, ensuring the model's response matches your Swift structs exactly for easier data handling.

Foundation Models for Apple AI

Name: Foundation Models for Apple AI
Author: johnrogers

byjohnrogers

•

Mobile Development

Implements on-device AI features like summarization and classification using Apple's Foundation Models framework in Swift.

About

This skill provides specialized guidance for integrating Apple's on-device Foundation Models framework into Swift applications, enabling privacy-first AI features without network dependencies. It assists developers in implementing text summarization, data extraction, and classification using Apple's 3B parameter language model. By leveraging patterns for @Generable type-safe outputs, Tool calling, and asynchronous streaming, this skill ensures developers can build responsive, intelligent iOS and macOS apps that respect user data privacy and platform constraints.

Key Features

Asynchronous streaming support for progressive UI updates
Type-safe structured output generation using the @Generable macro
Automated availability checks and error handling for Apple Intelligence
61 GitHub stars
Seamless tool calling for integrating MapKit, contacts, and external data
On-device AI integration for privacy-focused summarization and extraction

Use Cases

Extracting structured data from user text without server round-trips
Implementing real-time text classification within a SwiftUI application
Building a privacy-first note-taking app that summarizes content locally

About

Key Features

Asynchronous streaming support for progressive UI updates
Type-safe structured output generation using the @Generable macro
Automated availability checks and error handling for Apple Intelligence
61 GitHub stars
Seamless tool calling for integrating MapKit, contacts, and external data
On-device AI integration for privacy-focused summarization and extraction

Use Cases

Extracting structured data from user text without server round-trips
Implementing real-time text classification within a SwiftUI application
Building a privacy-first note-taking app that summarizes content locally