Can I run multiple requests concurrently on a single session?

No, LanguageModelSession handles one request at a time. You should check the isResponding property before sending a new request or create multiple sessions if needed.

What is the token limit for on-device models?

The Apple FoundationModels framework typically has a 4,096 token limit that includes the system instructions, user prompt, and the generated output combined.

Does this skill support offline AI generation?

Yes, the primary benefit of the FoundationModels framework is that it runs entirely on-device, supporting full offline functionality and ensuring no data leaves the device.

Which iOS version is required for these features?

These implementation patterns are specifically designed for the FoundationModels framework available in iOS 26 and later.

What is the @Generable macro used for?

The @Generable macro is used to generate structured Swift types instead of raw strings, providing compile-time safety and semantic guidance for the model to follow.

On-Device Apple Foundation Models

Name: On-Device Apple Foundation Models
Author: affaan-m

byaffaan-m

•

112,917

•

Mobile Development

Integrates Apple's on-device language models into iOS applications using the FoundationModels framework for privacy-first AI features.

This skill provides comprehensive patterns and implementation guides for leveraging Apple's native on-device LLMs within iOS 26+ environments. It enables developers to implement secure, offline-capable AI features including structured data extraction via the @Generable macro, custom tool calling for domain-specific tasks, and real-time UI updates through snapshot streaming. By focusing on Apple Intelligence integration, it helps developers build responsive, privacy-centric applications that process natural language without relying on external cloud services or data transfer.

Key Features

01Custom tool calling for domain-specific AI actions and logic

02Type-safe structured output generation using the @Generable macro

03Privacy-preserving on-device text generation and summarization

04Real-time UI updates via snapshot streaming of partial responses

05Built-in model availability and device eligibility status handling

06112,917 GitHub stars

Use Cases

01Extracting structured data like calendar events or profiles from natural language input

02Building privacy-sensitive applications that process user data without cloud dependency

03Developing AI-powered features that function entirely offline for travel or utility apps

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add affaan-m/everything-claude-code foundation-models-on-device

For use in Claude.ai and ChatGPT

Key Features

01Custom tool calling for domain-specific AI actions and logic

02Type-safe structured output generation using the @Generable macro

03Privacy-preserving on-device text generation and summarization

04Real-time UI updates via snapshot streaming of partial responses

05Built-in model availability and device eligibility status handling

06112,917 GitHub stars

Use Cases

01Extracting structured data like calendar events or profiles from natural language input

02Building privacy-sensitive applications that process user data without cloud dependency

03Developing AI-powered features that function entirely offline for travel or utility apps

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add affaan-m/everything-claude-code foundation-models-on-device

For use in Claude.ai and ChatGPT