Table of Contents
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
Product
Feature specifications, user flows, and pricing details for WardrobeAI.
Core Features
1. Style Interview
A conversational onboarding flow that replaces the typical multiple-choice style quiz.
Current MVP: Simple multi-step form (3 steps) Future: AI-powered conversational interview via LLM
Step 1: Who Are You
- Name
- Body type (hourglass, pear, apple, rectangle)
- Height
- Clothing size
Step 2: Style Preferences
- Primary occasion (work, casual, formal, evening)
- Preferred styles (select from: Professional, Casual, Elegant, Bohemian, Classic, Edgy, Preppy, Modern)
- Excluded styles (same choices, inverted selection)
Step 3: Colors & Budget
- Monthly budget slider ($50 - $500)
- Preferred colors (select from: Black, White, Blue, Red, Green, Purple, Gray, Brown)
- Seasons (Spring, Summer, Fall, Winter)
Frontend: frontend/pages/style-interview.js
Data stored in: style_profiles and user_profiles tables
2. Curated Box Generation
The core product experience — WardrobeAI selects items based on your profile and budget.
How it works:
- User selects or creates a style profile
- Backend runs matching algorithm against inventory
- Greedy selection picks items within budget
- Box is displayed with match scores and reasons
Algorithm Weights:
| Factor | Weight | Description |
|---|---|---|
| Style Match | 35% | Item styles overlap with profile preferences |
| Color Match | 35% | Item colors match preferred colors (with similarity grouping) |
| Budget Fit | 20% | Item price within budget (scored 1.0 at free → 0.5 at max) |
| Category Match | 10% | Item category matches must-have categories |
Scoring: Items scored 0-100. Items with score > 0 pass the filter. Top N items sorted descending, selected greedily until budget exhausted.
Frontend: frontend/pages/curated-box.js
3. Feedback Loop
Structured feedback that improves future curation.
Per-item feedback:
- Rating (1-5 stars)
- Keep / Return decision
- Optional notes
System learning:
- Tracks keep rates and average ratings per item
- Computing improvement scores for feedback analysis
- Future iterations can use this data to tune algorithm weights
Frontend Feedback Form:
// Each item card shows:
<select> // Rating: 1-5
<button> // Keep
<button> // Return
<textarea> // Notes (optional)
Improvement Score Formula:
improvementScore = (avgRating × 20) + (kept_count × 10) - (returned_count × 20)
4. Invariants System (Planned)
"Hard lines" — fashion dealbreakers that are always respected.
Examples:
- "No florals" → all floral items excluded
- "No synthetics" → only natural fibers
- "No crop tops" → excluded regardless of style match
- "Long sleeves only" → short sleeves filtered out
Status: Spec defined but not implemented in MVP codebase.
User Flows
Style Form → Box → Feedback
[Style Interview] → [Profile Saved] → [Generate Box] → [Review Items] → [Submit Feedback]
│ │ │ │ │
▼ ▼ ▼ ▼ ▼
Enter: POST /profiles POST /profiles/ View item cards POST /feedback
- Name (not in current :id/box with match scores
- Body type API - would need and prices
- Height to be added)
- Size
- Occasion
- Styles/colors
- Budget
The current MVP requires profiles to be pre-seeded via POST /seed. A production version would allow user-created profiles.
Pricing Tiers
Essential — $100/mo
- 3-5 clothing items per month
- Sourced from small businesses and boutique brands
- Style interview and feedback loop included
- Shoes, bags, jewelry available as add-ons
Signature — $250/mo
- 4-8 clothing items per month
- Higher-end boutique brands
- Early access to new designers
- Concierge-level style interview
Custom — Variable
- You set the budget
- You choose components (clothing only, or full wardrobe)
- Custom frequency (monthly, quarterly, etc.)
- Priority access to AI design-on-demand (Phase 3+)
MVP Phase: Only Essential tier active. No Stripe integration yet — handled manually.
Competitive Positioning
| Company | Model | Our Edge |
|---|---|---|
| Stitch Fix ($1.6B cap) | AI + stylist, mass brands | Small business sourcing; budget caps |
| Wantable | Curated edits, $20 fee | Conversational interview; feedback that learns |
| DailyLook | Premium styling, $40 fee | Starts at $100/mo; hard constraints honored |
| Rent the Runway | Rental subscription | Keep what you love; not rental |
| Dia & Co | Plus size niche | Size-inclusive, not niche-restricted |
The niche we own: Small business curation with budget caps and real learning.
Technical Architecture
Current (MVP)
Frontend (Next.js :3003) ──HTTP──→ Backend (Express :3002) ──SQL──→ SQLite
│ │
└── CSS Modules └── wardrobe-ai.db
└── Pages: index, style-interview, curated-box
Future (Phase 2+)
Static Site (Astro SSG) ──→ Storj
│
└── Style Interview
└── Landing + Waitlist
└── Brand pages
Chat Service ──→ Railway/Fly.io
│
├── Curation API
├── User management
├── Stripe billing
└── Feedback analytics
AI Service (Python) ──→ Railway/Fly.io
│
├── ML matching engine
├── LLM style interview
└── Design-on-demand
See the Architecture page for more detail.