The Wearly Intelligent Book¶
Wearly's Intelligent Book is the learning and reference companion for the styling-agent prototype. The app shows what Wearly does; the book explains how and why it reasons that way.
This is an intelligent textbook about AI-powered personal styling, using Wearly as the working case study — not app documentation, not a generic styling guide. Inspired by intelligent-textbooks.
What this book is — and how it relates to the app¶
Wearly ships as seven coordinated layers. The app is one of them; the book is the reference layer that explains every other one.
| Layer | What it is | Where it lives |
|---|---|---|
| App (prototype) | The agent running end-to-end on a real wardrobe. | streamlit run app.py |
| Book chapters | The reasoning behind every step the app takes. | book/01-vision.md … book/12-evidence-and-references.md |
| Skill rules | The operational rules the agent applies at runtime. | skills/wearly-styling-agent/ |
| Evidence & references | The verified sources the rules trace back to. | docs/evidence-and-references.md, References |
| Wearly Knowledge Graph | A three-layer structured-knowledge graph — domain rules + user-behavior aggregates + runtime archetypes. 91 nodes, 187 typed edges. | Wearly Knowledge Graph |
| MicroSims | Interactive widgets that let the reader try one rule or concept hands-on. | MicroSims |
A reader who only opens the app sees outcomes. A reader who only opens the book sees structure. Reading them together is the intelligent-textbook pattern: the prototype demonstrates the agent in motion; the book explains every line of its reasoning.
A Level-2 intelligent textbook¶
Wearly is built as a Level 2 intelligent textbook per Dan McCreary's framework: a concept-graph–driven, MicroSim-equipped, evidence-citable companion to the running prototype. That framing is concrete — Wearly ships the artifacts a Level-2 book is expected to ship:
- a Wearly Knowledge Graph — a three-layer structured-knowledge graph (domain rules, user behavior, runtime archetypes), 91 nodes, 187 typed edges, deterministically generated from public seed data;
- MicroSims — small interactive simulations for color harmony, freshness, and the knowledge graph;
- a project-wide Glossary, FAQ, and References;
- per-chapter Key terms and Self-check footers (Bloom-tiered);
- a canonical
book-metadata.ymlat the repo root.
See book-metadata.yml for the canonical
version, license, and framework pointers.
Table of contents¶
The book is organized in three parts. Part I explains the agent — how it reasons, step by step. Part II is the Styling Rule Reference — the styling knowledge the agent applies, taught learner-first with evidence. Part III is the evidence base itself.
Part I — The Wearly Agent¶
| # | Chapter | What it covers |
|---|---|---|
| 01 | Product vision | What Wearly is, who it's for, what it isn't |
| 02 | User problem | The 15-minute closet decision and why it matters |
| 03 | Agent workflow | The seven-step reasoning loop, end to end |
| 04 | Styling knowledge base | How the rule packs are organized — entry point to Part II |
| 05 | Wardrobe intelligence | How wardrobe items are modeled, filtered, and matched |
| 06 | Fit-profile logic | Body-positive personalization, what's modeled today and what's next |
| 07 | Calendar & weather context | Two external context sources and how they steer decisions |
| 08 | Shopping-gap logic | When the agent says "you don't own this yet" |
| 09 | Before / after demo | The delta between manual planning and an agentic system |
| 10 | Privacy & security | Data handling, today and in production |
| 11 | Product roadmap | Where Wearly is going |
Part II — Styling Rule Reference¶
Each chapter teaches one piece of styling knowledge the agent reasons over, cites the evidence behind it, and points at the rule pack and Part-I chapter that consume it.
| # | Chapter | What it covers |
|---|---|---|
| S1 | Color and skin tone | The warm/cool axis, three palettes, the +8 / +4 / −10 scoring formula |
| S2 | Silhouette and fit | Proportion-related styling as user-declared preferences, never classification |
| S3 | Occasion and context | The five canonical tags and why this is the minimum useful taxonomy |
| S4 | Weather and layering | Temperature bands, the <60°F outerwear threshold, the seasonal fallback |
| S5 | Wardrobe construction | Required pieces, dress-vs-separates branching, the accessory cap |
| S6 | Rotation and freshness | Wear-history models, the 0.4 floor, "let it breathe" |
| S7 | Honest gaps | Qualified gaps, descriptive-not-promotional tone, honest gaps over forced fits |
| S8 | Body-positive framing as a discipline | The forbidden-token contract enforced in code, not only in prose |
Part III — Evidence & provenance¶
| # | Chapter | What it covers |
|---|---|---|
| 12 | Evidence and references | What informs Wearly's recommendation logic, and what still needs citation work |
How to read this book¶
- In order for the full story.
- By chapter if you're verifying a specific claim in the app.
- As a reference if you're building on the system — every chapter names the file or data structure it documents.
Each chapter is short. The principle is structure beats volume: a small set of organized chapters that match the system, rather than a long unstructured prose.
Companion to Wearly_Product_Brief.md (the canonical product vision) and docs/architecture.md (the system snapshot).