What problem does it solve?
This Skill eliminates the need to send user data to cloud servers for AI processing, solving privacy risks and offline functionality gaps for iOS app AI features.
Core Features & Use Cases
- On-device text generation: Build AI-powered features like content summarization, conversational assistants, and text creation that run entirely on the user's device with no cloud dependency.
- Structured data extraction: Use the @Generable macro to pull organized, typed data (like form inputs, event details) from natural language instead of parsing error-prone raw text strings.
- Custom tool calling: Create domain-specific AI actions (like recipe search, calendar lookups) that the model can invoke automatically based on user prompts.
- Real-time response streaming: Use snapshot streaming to progressively display generated structured content in SwiftUI for smooth, responsive user interfaces.
A common use case is building a privacy-focused cooking app that suggests recipes based on user-provided ingredients, works fully offline, and returns structured recipe data without sending any user input to external servers.
Quick Start
Use the foundation-models-on-device skill to add an on-device AI feature to your iOS app that extracts structured cat profile information (name, age, personality description) from natural language user input using the @Generable macro, with no data leaving the user's device.
