From journals to gym routines: Apple’s foundation models put AI in everyday apps

Apple foundational models

Apple’s big AI bet is starting to pay off. With iOS 26, iPadOS 26 and macOS 26 now rolling out, developers have access to the Foundation Models framework, the engine that powers Apple Intelligence. The pitch is simple: on-device large language models, free to use, private by design, and available offline.

According to Apple, developers have already begun weaving these tools into apps that millions rely on — from fitness trackers to journaling platforms.

Smarter Health and Fitness

Apps like SmartGym are showcasing what’s possible. A user can now describe a workout in plain language and see it converted into a structured routine — complete with reps, rest times and equipment tweaks. SmartGym’s “Smart Trainer” also explains the reasoning behind its recommendations, so the coaching feels less robotic and more human.

Other health apps are also joining in. SwingVision uses Apple’s framework to analyse tennis videos and generate precise feedback, while 7 Minute Workout builds personalised routines that avoid injuries or prep users for events. Even journaling apps like Gratitude are generating weekly reflections and affirmations on-device.

Rethinking Journaling and Reflection

Privacy is central to Apple’s pitch, and journaling app Stoic is an example the company is showcasing. It generates personalised prompts, contextual reminders, and even summaries of past entries — all without data leaving the device. Its developer notes that features which once required heavy back-end infrastructure can now run natively, making personalised mental-health support more accessible and secure.

Education Apps Get Interactive

Learning apps are also tapping in. CellWalk, a biology app, now allows students to click on complex scientific terms and receive conversational explanations tailored to their knowledge level. Similarly, Grammo’s AI tutor generates new grammar questions on the fly, while Lil Artist creates illustrated stories for children using structured prompts rather than open-ended text, making AI approachable for younger learners.

Productivity and Creativity Boost

The productivity space is buzzing too. Task manager Stuff now supports natural language input (“Call Sophia Friday”) and can even scan handwritten notes into to-dos. Video editing app VLLO analyses clips and automatically suggests background music and stickers, while Detail: AI Video Editor can turn rough outlines into polished scripts. Document-heavy apps like Signeasy and Essayist now offer instant summaries, references and conversational search.

Building with Foundation Models

For developers, the framework is tightly integrated with Swift and supports guided generation for predictable outputs. Models can also call back into apps for more context, ensuring responses remain accurate and grounded.

By making on-device intelligence free and developer-friendly, Apple is betting its framework will quietly reshape apps across health, education and productivity.