Intelligence Catalog is an interactive knowledge base and playground for on-device AI — discover, learn, and experience real models running through curated demos and educational content, right on your iPhone, iPad, and Mac.
Designed to feel like a mix of Apple sample code, Apple Developer Documentation, and an interactive AI museum.
Explore a curated catalog of Apple's Core ML and Vision models.
Try each model yourself with your own photos and text.
See inputs, outputs, and how-to-use tips for every model.
Every demo runs locally — nothing you provide ever leaves your device.
A real capture of the app's first-launch experience, next to the Gallery and Model Detail screens you'll land on right after.
Real, bundled Core ML models alongside native Vision and Natural Language tasks and Apple Intelligence–powered Foundation Models — spanning image and text understanding.
FastViT, MobileNetV2, ResNet-50, YOLOv3, DETR, DeepLabv3, Depth Anything V2, MNIST, and more.
Face, barcode, animal & pose detection, text recognition, saliency, document scanning, and more.
Apple Intelligence–powered summarization and structured data extraction.
Sentiment analysis, tokenization, language ID, named entity recognition, and part-of-speech tagging.
A dedicated Learn tab pairs every topic with a Practical track (real code — Keras next to Apple's own on-device equivalent) and a Theory track (how it actually works, with a diagram where one helps) — then quizzes you on it.
Every Practical write-up ends with a Keras/Python snippet next to Apple's actual on-device API call — syntax-highlighted, matching what the app's own demo really runs.
Hand-drawn, theme-aware diagrams — a single neuron, a CNN pipeline, self-attention, IoU, BIO tagging, precision/recall — for the topics where a picture beats a paragraph.
Each topic draws a fresh, randomized set from a 50-question bank every session, and a local Progress panel tracks your score by topic — no account, no server.
"An interactive playground for Apple's on-device AI — learn, explore, run, and understand, all in one native app."
Built natively in SwiftUI for iOS, iPadOS & Mac · Fully on-device · Dark mode supported