---
title: "The official catalog: recipes and registry"
lesson_id: "01"
---

# The official catalog: recipes and registry

- **Lesson ID:** 01
- **Goal:** Explain Apple's official `apple/coreai-models` repository, its recipes, and the three typed registry groups.
- **Human lesson:** [01-the-official-catalog.html](01-the-official-catalog.html)

## Prerequisites

- A basic idea that an app feature takes an input and gives an output.
- The [model architectures course](../model-architectures-plain/) for language versus image shapes.

## Inputs, outputs, and artifacts

- **Inputs:** The official repository URL and the supplied typed registry catalog.
- **Outputs:** A simple difference between recipes, registry types, and a custom trained model.
- **Artifacts:** Progressive `.flow` diagrams, official repository link, and links to all three prior courses.

## Agent build steps

1. Name and link `https://github.com/apple/coreai-models` as Apple's official repository.
2. Explain `model_registry.py` as the source of the published typed presets for this course.
3. Teach the three groups in order: `llm`, `diffusion`, and `utility`.
4. Contrast a registry preset with ClothingCNN, which the reader trained in the forward course.
5. Mention that `vlm` and `parakeet` folders are visible in the tree but do not invent model names for unlisted registry presets.
6. Link the architectures course and From the metal to the model, then point to lesson 02.

## Constraints

Use only supplied catalog facts. Keep the explanation simple and friendly. No math, CUDA, quizzes, Mermaid, or em dashes.

## Key concepts

- Start at `https://github.com/apple/coreai-models` for recipes and the typed list.
- `model_registry.py` is the source of published presets for this course.
- The three groups are `llm`, `diffusion`, and `utility`.
- Type tells you the kind of output. It does not promise the same speed or device support.
- `vlm` and `parakeet` folders can appear in the tree. They are not typed catalog presets here.

## Takeaways

- You can name the official repo as the source of recipes and registry types.
- You can describe `llm`, `diffusion`, and `utility` by output shape.
- You can choose input and output first, then type, then a named preset.

## Acceptance checks

- The official repository URL is linked.
- `llm`, `diffusion`, and `utility` are described as the three groups.
- ClothingCNN is clearly a custom trained contrast.
- `vlm` and `parakeet` are mentioned only as additional folders without invented names.
- Progressive `.flow` diagrams and the Agent brief link are present.