WARDROBE AI

WARDROBE AI is an AI wardrobe recommendation PWA that combines a Node.js / Express backend, PostgreSQL on Supabase, JWT auth, and a rule-based styling engine for color season, occasion, silhouette, material, and seasonal outfit guidance.

Live Demo

• PWA Demo: https://wardrobe-ai-pwa-v2.onrender.com
• API Test: https://wardrobe-ai-pwa-v2.onrender.com/api/test
• Database Health: https://wardrobe-ai-pwa-v2.onrender.com/api/health/db
• Mobile Demo Checklist: docs/MOBILE_DEMO_CHECKLIST.md

Render Free services may cold start after inactivity. For interviews, open the demo 5-10 minutes before presenting.

The project was repositioned from an APK-only mobile demo into a phone-friendly PWA, so it can be opened directly from iPhone Safari, Android Chrome, or a desktop browser during interviews and portfolio reviews.

Current Status

• Production PWA is deployed on Render.
• Production database is deployed on Supabase.
• Render health checks and database-backed endpoints are verified.
• Mobile API base URL points to the new production service.
• GitHub README, release notes, mobile demo checklist, and infrastructure notes are updated for portfolio use.

Verified endpoints:

/                         200
/api/test                 200
/api/health/config        200
/api/health/db            200
/api/meta/seasons         200
/api/meta/occasions       200
/api/meta/materials       200

Core Features

• AI outfit recommendation flow for occasion and season.
• 12 color-season reference system.
• Wardrobe item attributes for category, color, material, silhouette, pattern, drape, style tags, and season tags.
• Personalized profile inputs for color season, skin tone, and body type.
• Outfit history and feedback loop for future recommendation tuning.
• PWA deployment for phone-based live demonstration.

Screenshots

Outfit Recommendation

Cold Start Color Guide

Personal Settings

Wardrobe State

Tech Stack

• Frontend: Expo, React Native, TypeScript, Expo Web / PWA
• Backend: Node.js, Express
• Database: PostgreSQL / Supabase
• Auth: JWT access token and refresh token
• Deployment: Render
• Container: Docker / docker-compose
• Testing: Node-based backend verification and production endpoint checks

API Example

Create an account and get a token:

curl -X POST https://wardrobe-ai-pwa-v2.onrender.com/api/auth/register \
  -H "Content-Type: application/json" \
  -d '{"email":"[email protected]","password":"demo-password-123"}'

Recommendation request:

POST /api/recommend
Authorization: Bearer <access_token>
Content-Type: application/json

{
  "occasion": "work_daily",
  "season": "summer",
  "top_n": 3
}

Example response shape:

{
  "recommendations": [
    {
      "score": 8.7,
      "outfit": {
        "layers": [],
        "bottom": null,
        "accessories": {}
      },
      "analysis": ["color harmony", "occasion fit", "season fit"]
    }
  ]
}

Local Development

npm install
npm start

Create a local environment file from the example:

cp .env.example .env

PWA build:

cd mobile
npm install
npm run build:web

The server serves both the API and the generated PWA static files.

Docker

Build and run the app container:

docker build -t wardrobe-ai .
docker run --env-file .env -p 3000:3000 wardrobe-ai

Run app + local PostgreSQL:

docker compose up --build

The compose file starts PostgreSQL for local development. Apply the database schema from db/ before testing database-backed endpoints locally.

Production Verification

npm run verify:production

Manual production checks:

https://wardrobe-ai-pwa-v2.onrender.com/
https://wardrobe-ai-pwa-v2.onrender.com/api/test
https://wardrobe-ai-pwa-v2.onrender.com/api/health/config
https://wardrobe-ai-pwa-v2.onrender.com/api/health/db
https://wardrobe-ai-pwa-v2.onrender.com/api/meta/seasons
https://wardrobe-ai-pwa-v2.onrender.com/api/meta/occasions
https://wardrobe-ai-pwa-v2.onrender.com/api/meta/materials

Portfolio Notes

WARDROBE AI is the main full-stack AI product in the 30-day GitHub portfolio plan. The project now satisfies the interview demo requirement: it can be opened from a phone as a live, hosted, database-backed PWA.