How Developers Are Utilizing Apple’s Local AI Models in iOS 26
Earlier this year, Apple unveiled its Foundation Models framework during WWDC 2025, enabling developers to leverage the company’s local AI models to enhance features in their applications.
The company highlighted that this framework grants developers access to AI models without concerns about inference costs. Additionally, these local models come equipped with capabilities like guided generation and tool calling.
As iOS 26 rolls out to all users, developers have been refreshing their apps to incorporate features fueled by Apple’s local AI models. Although Apple’s models are smaller compared to leading models from OpenAI, Anthropic, Google, or Meta, the local-only features primarily enhance the quality of life within these apps rather than bringing significant changes to the app’s workflow.
Below are some of the first apps to harness Apple’s AI framework.
Lil Artist
The Lil Artist app provides various interactive experiences that assist children in learning skills like creativity, math, and music. Developers Arima Jain and Aman Jain launched an AI story creator with the iOS 26 update, allowing users to select a character and theme, with the app generating a story using AI. The text generation for the story is powered by the local model.
Daylish
The developer of the Daylish app is creating a prototype for automatically suggesting emojis for timeline events based on the title in the daily planner app.
MoneyCoach
The finance tracking app MoneyCoach includes two clever features powered by local models. First, the app provides insights regarding spending, such as whether users spent more than average on groceries for a specific week. The second feature automatically suggests categories and subcategories for spending items for swift entries.
LookUp
The word learning app LookUp has implemented two new modes utilizing Apple’s AI models. One new learning mode employs a local model to generate examples relevant to a word and prompts users to explain the word’s usage in a sentence.
The developer is also leveraging on-device models to create a map view displaying a word’s origin.
Tasks
Similar to other apps, the Tasks app has integrated a feature that automatically suggests tags for an entry using local models. It also utilizes these models to detect and schedule recurring tasks. Users can verbally input items, and the local model will break them down into various tasks without requiring internet access.
Day One
The journaling app Day One, owned by Automattic, employs Apple’s models to highlight and suggest titles for entries. The team has also introduced a feature that generates prompts encouraging users to elaborate and write more based on their previous entries.
Crouton
The recipe app Crouton leverages Apple Intelligence to suggest tags for recipes and assign names to timers. It also employs AI to break down lengthy text into simple, step-by-step cooking instructions.
Signeasy
The digital signing app Signeasy utilizes Apple’s local models to extract key insights from contracts and provide users with summaries of the documents they are signing.
Dark Noise
The background sound app Dark Noise allows users to describe a soundscape in a few words, generating one based on their input. Users can adjust various elements of the soundscape once created.
Lights Out
Lights Out is a newly launched app tracking the F1 season and grand prix, developed by Shihab Mehboob, the creator of the Twitter client Avery and Mastodon client Mammoth. The app relies on on-device AI models to summarize commentary during races.
Capture
The note-taking app Capture employs local AI to provide category suggestions to users as they type in their notes or tasks.
Lumy
The sun and weather tracking app Lumy now offers smart weather-related suggestions using AI.
CardPointers
CardPointers is an app designed to help users track credit card expenses and optimize point earnings. The app’s latest version employs AI to allow users to ask questions about their cards and offers.
Guitar Wiz
The guitar learning app Guitar Wiz utilizes the Apple Foundation Model framework in several ways. Users receive explanations for chords while learning and advanced insights based on time intervals. Additionally, the AI model supports over 15 languages to assist developers.
SmartGym
The SmartGym app uses local AI to transform workout descriptions into step-by-step lists with rep counts, intervals, and equipment. It also provides users with workout summaries, monthly progress, and detailed breakdowns of individual exercise performance.
Stoic
The journaling app Stoic incorporates Apple’s models to provide personalized prompts based on user mood logging. These models can also assist in summarizing posts, searching past entries, and organizing them.
SwingVision
This app aids players of racquet sports like tennis and pickleball in enhancing their form based on video recordings. The developers now utilize foundational models to offer actionable and precise feedback.
Zoho
The India-based productivity suite Zoho employs local models to enhance summarization, translation, and transcription across its applications, including Notebook for documents and Tables for spreadsheets.
TrainFitness
The workout app uses on-device models to propose alternatives for exercises when specific equipment is unavailable.
Stuff
The to-do app Stuff features a voice mode that leverages Apple’s AI models to listen to the user and convert spoken input into individual tasks.
We will continue to update this list as we discover more apps utilizing Apple’s local models.
