Apple’s introduction of the Foundation Models framework at WWDC 2025 has enabled developers to integrate local AI capabilities directly into iOS apps, now fully available with the rollout of iOS 26. These on-device models, while smaller than those from OpenAI, Google, or Meta, offer powerful features such as guided generation and tool calling without any inference costs or reliance on internet connectivity. The shift toward local AI is particularly beneficial for privacy-focused applications and features that enhance user experience through subtle but meaningful improvements. Here are some of the first apps leveraging Apple’s local AI models in iOS 26: Lil Artist, a learning app for children, now includes an AI-powered story creator. Users can pick a character and theme, and the app generates a custom story using the local model for text creation. Daylish is testing a prototype that automatically suggests emojis for daily planner events based on the event title, streamlining user input. MoneyCoach, a finance tracker, uses local AI to provide spending insights—like identifying if a user spent more than usual on groceries—and to suggest appropriate categories and subcategories for quick expense entry. LookUp, a vocabulary app, has added a new learning mode that generates real-world examples for words and prompts users to explain their usage in sentences. It also uses on-device models to create visual maps of a word’s etymological origins. Tasks introduces AI-driven tag suggestions, detects recurring tasks for automatic scheduling, and allows users to speak their to-do items, which the local model then breaks down into structured tasks—all without going online. Day One, the journaling app, now uses Apple’s AI to generate entry highlights, suggest titles, and create follow-up prompts that encourage deeper reflection based on what the user has already written. Crouton, a recipe app, leverages local AI to suggest tags for recipes, name timers, and convert long blocks of text into clear, step-by-step cooking instructions. SignEasy uses on-device models to extract key information from contracts and deliver concise summaries before signing. Dark Noise, a background sound app, lets users describe a desired soundscape in plain language and generates a custom mix. Users can then fine-tune individual sound elements. Lights Out, a Formula 1 tracking app, uses local AI to summarize live race commentary, offering real-time insights during races. Capture, a note-taking app, provides category suggestions as users type, improving organization without cloud dependency. Lumy, a sun and weather app, now delivers AI-powered weather-related tips and recommendations based on real-time conditions. Cardpointers, a credit card rewards tracker, allows users to ask questions about their cards and rewards programs, with AI delivering instant, on-device responses. Guitar Wiz, a music learning app, uses local models to explain chords, offer advanced playing insights based on timing, and supports over 15 languages to reach a broader audience. These early adopters demonstrate how Apple’s local AI framework is being used to enhance functionality, improve usability, and maintain privacy—all within the constraints of on-device processing. As more developers explore the potential of Foundation Models, we can expect even more innovative, intelligent, and responsive features across the iOS ecosystem.