Apple Foundation Models framework
Also known as: Foundation Models framework, AFM framework, Apple FM framework, Apple Intelligence SDK
The Apple Foundation Models framework is a Swift API that lets any developer tap into the same AI models powering Apple Intelligence, without needing API keys, cloud costs, or internet access for on-device tasks. The on-device model handles language understanding, structured output, and tool calling. Developers write a few lines of Swift, create a session, and get responses back, with no provider sign-up required.
At WWDC 2026, Apple expanded the framework significantly. It now supports image input alongside text, so apps can analyze photos, screenshots, and visual content on-device. A new LanguageModel protocol opens the framework to nearly any model: Apple's own on-device model, Private Cloud Compute models, or server-side models from third parties. Anthropic and Google are both shipping Swift packages so developers can call Claude and Gemini through the same interface they already use for on-device inference. A new Dynamic Profiles system lets an app swap models, tools, and instructions mid-session, which Apple is positioning as the foundation for multi-agent workflows inside iOS apps.
For qualifying small developers enrolled in the App Store Small Business Program with fewer than two million total first-time downloads, Apple is offering free access to its cloud models on Private Cloud Compute with no per-token cost. The framework is also going open source in summer 2026. For iOS builders specifically, this is one of the most significant developer surface changes in years: a single Swift API can now route between a private on-device model, a free Apple cloud tier, or a frontier model like Claude, chosen per request based on cost, latency, and privacy needs.