apple-on-device-ai-macos

What problem does it solve?

This Skill eliminates the complexity of building performant, private on-device AI and machine learning features for macOS 26+ apps, removing the need for cloud services and handling low-level integration with Apple's native ML frameworks and open-weight inference tools.

Core Features & Use Cases

• Apple Foundation Models Integration: Use SystemLanguageModel for on-device text generation, structured output with @Generable, and multi-turn tool calling.
• CoreML & MLTensor Workflows: Configure Neural Engine/GPU/CPU compute units, build lazy data pipelines, and deploy models via app bundle or on-demand resources.
• Production-Grade Inference: Isolate heavy model workloads in XPC helper processes to prevent main app crashes, and use MLX Swift or llama.cpp for open-weight LLM deployment.
• Use Case: Build a macOS writing assistant that uses Foundation Models for real-time text summarization, CoreML for custom grammar checking, and XPC isolation to keep the UI responsive during inference.

Quick Start

Use the apple-on-device-ai-macos skill to add private on-device text generation with streaming responses to your macOS app using Apple's Foundation Models framework.