Can I run multiple AI requests simultaneously?

A single session can only handle one request at a time (check the isResponding property). For concurrent requests, you must initialize separate LanguageModelSession instances.

What are the benefits of using on-device Foundation Models?

The primary benefits are enhanced user privacy, as data never leaves the device, and the ability for AI features to function without an internet connection.

What iOS version is required for this skill?

These implementation patterns are designed for iOS 26 and later, which supports the Apple FoundationModels framework.

How does @Generable improve AI integration?

The @Generable macro allows developers to define structured Swift types that the model must follow, providing compile-time safety and eliminating the need for manual string parsing.

What is the context window for Apple's on-device models?

The framework currently has a 4,096 token limit, which includes the combination of system instructions, user prompts, and the generated output.

On-Device Apple Foundation Models

Name: On-Device Apple Foundation Models
Author: Mark393295827

byMark393295827

•

Mobile Development

Integrates Apple's on-device FoundationModels framework for privacy-preserving, offline LLM capabilities in iOS applications.

This skill provides specialized patterns and best practices for leveraging Apple's FoundationModels framework in iOS 26+ environments. It enables developers to integrate on-device language models for text generation, structured data extraction using the @Generable macro, and custom tool calling—all while ensuring user data remains on-device for maximum privacy and offline availability. By following these implementation patterns, developers can handle model availability checks, manage multi-turn sessions, and implement real-time snapshot streaming for responsive, AI-powered mobile interfaces.

主な機能

01Guided generation with @Generable for type-safe structured output

02On-device text generation and summarization for offline support

03Real-time snapshot streaming for progressive UI updates

04Custom tool calling to connect local app logic with AI actions

051 GitHub stars

06Advanced session management with built-in availability checks

ユースケース

01Building privacy-centric AI assistants that process sensitive data locally

02Extracting structured data from natural language input to populate Swift forms

03Developing AI features that require high performance and zero latency without cloud dependencies

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mark393295827/house-maint-ai foundation-models-on-device

For use in Claude.ai and ChatGPT

主な機能

01Guided generation with @Generable for type-safe structured output

02On-device text generation and summarization for offline support

03Real-time snapshot streaming for progressive UI updates

04Custom tool calling to connect local app logic with AI actions

051 GitHub stars

06Advanced session management with built-in availability checks

ユースケース

01Building privacy-centric AI assistants that process sensitive data locally

02Extracting structured data from natural language input to populate Swift forms

03Developing AI features that require high performance and zero latency without cloud dependencies

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mark393295827/house-maint-ai foundation-models-on-device

For use in Claude.ai and ChatGPT