Does this include WWDC 2025 code samples?

Yes, the skill includes specific code examples and patterns sourced directly from WWDC 2025 sessions 286 and 301.

What platforms does this Foundation Models reference cover?

This skill covers modern xOS development including iOS 26+, macOS 26+, iPadOS 26+, and visionOS 26+.

Does the Foundation Models framework require an internet connection?

No, this reference is for Apple's on-device models which run entirely offline, ensuring user privacy and zero per-request costs.

What is the difference between @Generable and @Guide?

@Generable is used to define the structure of the output, while @Guide provides specific constraints like numeric ranges, array counts, or regex patterns to the model.

Can this skill help me with structured JSON output?

Yes, it provides comprehensive documentation on using the @Generable macro and constrained decoding to ensure AI responses match your Swift structs perfectly.

Foundation Models Reference

Name: Foundation Models Reference
Author: CharlesWiltgen

byCharlesWiltgen

•

233

Mobile Development

Provides a comprehensive technical reference for Apple's on-device Foundation Models framework using Swift and WWDC 2025 patterns.

About

The axiom-foundation-models-ref skill is a specialized technical documentation repository for developers building privacy-first AI features on Apple platforms. It offers an exhaustive guide to the Foundation Models framework, covering essential components like LanguageModelSession, structured output with the @Generable macro, and constrained decoding using @Guide. By leveraging on-device 3B parameter models, this skill helps developers implement text summarization, content classification, and streaming interactions directly within iOS, macOS, and visionOS applications without external API costs or network dependencies.

Key Features

Technical specifications for Apple's 3B parameter on-device LLM
233 GitHub stars
Detailed implementation patterns for the @Generable and @Guide macros
Constrained decoding guides for guaranteeing structural correctness of AI output
Complete API reference for LanguageModelSession and structured data generation
Snapshot streaming documentation for real-time SwiftUI updates

Use Cases

Generating type-safe, structured Swift objects from on-device AI prompts
Architecting multi-turn AI conversations with maintained transcript history
Implementing offline-capable text summarization and entity detection in mobile apps

About

Key Features

Technical specifications for Apple's 3B parameter on-device LLM
233 GitHub stars
Detailed implementation patterns for the @Generable and @Guide macros
Constrained decoding guides for guaranteeing structural correctness of AI output
Complete API reference for LanguageModelSession and structured data generation
Snapshot streaming documentation for real-time SwiftUI updates

Use Cases

Generating type-safe, structured Swift objects from on-device AI prompts
Architecting multi-turn AI conversations with maintained transcript history
Implementing offline-capable text summarization and entity detection in mobile apps