Apple Unveils Foundation Models Framework, Letting Developers Build Native AI Experiences with Apple Intelligence

At WWDC 2025, Apple made one of its most significant announcements yet in the AI space: developers will soon have direct access to the same powerful language models that drive Apple Intelligence through a new tool called the Foundation Models framework. This move marks a pivotal shift in Apple’s AI strategy, opening the door for third-party developers to create deeply integrated, privacy-first generative AI experiences directly within their apps — all without relying on the cloud.

Rather than offering just surface-level features, the Foundation Models framework gives apps access to Apple’s on-device models for generating text, crafting images, summarizing content, and more. This capability enables developers to build AI-powered tools that run locally on iPhones, iPads, and Macs, making them faster, more secure, and less reliant on constant internet connectivity.

The framework represents a fundamental step in Apple’s approach to artificial intelligence — one focused on performance, battery efficiency, and user privacy. Unlike many competing platforms that send data to remote servers, Apple’s models are designed to stay on the device whenever possible, aligning with its long-standing emphasis on data protection.

Interestingly, Apple’s latest AI announcement not only confirms previous speculation but surpasses it. Earlier reports hinted that third-party apps might eventually get limited access to Apple Intelligence-powered features. However, Apple’s decision to open up the underlying models themselves — via a developer-accessible framework — suggests a more developer-centric vision than previously expected.

The introduction of Foundation Models also coincides with a broader AI upgrade across Apple platforms. Support for additional languages has been expanded, ensuring that the generative capabilities of Apple Intelligence can reach a wider global audience. Behind the scenes, Apple has been refining the performance and efficiency of its models, helping developers deliver responsive, native AI functionality without compromising device performance.

This development could have a transformative effect on how AI is experienced on Apple devices. From productivity tools and creative apps to health and education software, the ability to embed Apple’s system-level intelligence natively opens new possibilities for contextual, user-aware features — all handled privately and locally.

Apple's announcement also signals a quiet but clear challenge to other AI platforms like OpenAI and Google, which primarily operate through cloud APIs. With Apple prioritizing on-device execution, developers now face a compelling new option that combines performance with privacy, especially in regions where data sovereignty is a growing concern.

The Foundation Models framework is expected to launch later this year, giving developers the tools they need to craft more intelligent, efficient, and secure apps — all powered by the same models running inside Apple Intelligence.