foundation-models-on-device

Run an LLM on-device for iOS offline text generation by integrating Apple's FoundationModels framework, which enables privacy-preserving, low-latency AI features directly on the hardware without cloud reliance.

Get structured output from an on-device LLM in Swift by applying the @Generable macro, enforcing type-safe data extraction directly from the model session into defined Swift types.

Stream LLM generations to a SwiftUI interface in real-time using snapshot streaming, which provides continuous structured data updates to the UI as the model generates content.

Yes, you need iOS 26 or later to use Apple Intelligence FoundationModels, as the framework requires this platform version to support native on-device model sessions and tool execution.

Let an on-device LLM perform database lookups by defining custom Swift tools, allowing the model session to execute domain-specific calculations and actions securely during generation.

Build a privacy-first AI app without servers by leveraging Apple's FoundationModels framework, which processes data and executes tools entirely on-device to preserve user privacy.