Apple Foundation Models

VaultAudit — Tracks subscriptions by reading screenshots of receipts or Apple subscription pages on-device, without bank connections.

OmniNotch — Puts a multi-tool dashboard in the Mac notch with sales stats, AI usage tracking, a teleprompter, clipboard history, a file shelf, and music, weather, calendar and note widgets.

AI Coach — Runs a desktop pet on Mac that hatches, grows, and unlocks small tools, with optional links to Claude and website drop-ins that subscribe to RSS or save bookmarks.

ReturnNotice — Tracks product return windows and deadlines with countdowns, reminders, shipment details and refund status.

Tidy — Renames image and PDF files based on the text recognized inside them and sorts Downloads and Desktop with rules.

Organize My Life — Sorts tasks into Now, Next, and Someday buckets with a hard cap on the Now bucket and suggests buckets via local AI that the user must confirm.

Fog — Automatically names, groups and organizes notes on-device using Apple's Foundation Models, and answers questions about them.

Spotter — Logs strength training sets on iPhone and Apple Watch, storing all data on the device and syncing through the user's iCloud.

uChecker — Proofreads and rephrases selected text in any Mac app using on-device spelling checks, Apple's local model, or a user-chosen API endpoint.

StepGrab — Records a user's clicks on macOS and turns them into a numbered step-by-step guide with screenshots, exportable as PDF, GIF, or MP4.

Stacksmith — Manages native local development services defined in a stacksmith.yml file, with start/stop, logs, health checks, port conflict detection, and AI-assisted startup diagnosis.

Everwrite — Builds a story bible of characters, places and objects from a manuscript and flags continuity errors like name drift or contradictions between chapters.

macbook-ai — Wraps macOS built-in speech recognition, text-to-speech, and on-device language model into a Python library.

Tastik — Manages customizable lists with multiple formats, interactive widgets, iCloud sharing, templates, and on-device AI parsing of typed items.

Council — Sends one question to several AI models in parallel, has them critique each other, and shows where their answers agree and disagree along with a synthesis.

Foundation Eval Labs — Runs evals and traces against Apple Foundation Models and can be driven by an agent through MCP.