👕 AI Smart Wardrobe

Your Personal AI Stylist & Wardrobe Manager

Upload clothing photos, remove backgrounds automatically, classify garments with AI vision, and get weather-aware outfit recommendations — all in one app.

English | 简体中文 | 日本語

---

✨ Features

Feature	Description
Smart Upload	Upload clothing photos, auto-remove backgrounds with rembg, analyze category, color, and style via AI vision models
Weather-Based Styling	Integrates free global weather data (Open-Meteo) to generate outfit suggestions based on real-time conditions
Digital Wardrobe	Browse, search, and manage your clothing in a structured wardrobe view
AI Recommendations	Supports Gemini and OpenAI-compatible providers for personalized outfit generation
Recommendation Modes	Switch between balanced, goal_first, and wardrobe_first to match different styling priorities
Explainable Picks	Shows selection reasons for top/bottom/shoes so each outfit choice is traceable
Voice Goal Input	Use microphone input for scenarios like commute/date/sport to guide recommendation intent
Responsive UI	Optimized for desktop, tablet, and mobile with a modern Tailwind CSS interface

📸 Screenshots (New UI)

New Item Entry	Wardrobe View
AI Recommendation	Clothes Detail

🏗️ Tech Stack

Frontend	React + Vite + Tailwind CSS
Backend	FastAPI + SQLite
AI	Google Gemini / OpenAI-compatible APIs + rembg
Deploy	Docker / Docker Compose (amd64 & arm64)

🚀 Getting Started

Prerequisites

• Node.js v20+  |  Python v3.10+
• Google Gemini API Key or an OpenAI-compatible API key

1. Clone

git clone https://github.com/leoz9/AIWardrobe.git
cd AIWardrobe

2. Install Dependencies

# Backend
cd backend
python -m venv venv
source venv/bin/activate   # Windows: venv\Scripts\activate
pip install -r requirements.txt
cd ..

# Frontend
cd frontend && npm install && cd ..

3. Start

# One-command start (macOS/Linux)
chmod +x start.sh && ./start.sh

# Windows
start.bat

After startup:

• Frontend: http://localhost:5173
• Backend API: http://localhost:8000
• API Docs: http://localhost:8000/docs

Then open Settings in the UI to fill API Base/API Key/Model visually.

Manual start (separate terminals)

# Terminal 1: Backend
cd backend && source venv/bin/activate
uvicorn main:app --host 0.0.0.0 --reload --port 8000

# Terminal 2: Frontend
cd frontend && npm run dev

🧠 Recommendation API (Modes, Goal, Explainability)

GET /api/recommendation

Common query params:

• location: city string or coordinates
• mode: balanced | goal_first | wardrobe_first
• goal: optional scenario, e.g. commute, date, sport, interview

Example:

curl "http://localhost:8000/api/recommendation?location=Shanghai,Shanghai,China&mode=goal_first&goal=commute"

Response highlights:

• mode: applied recommendation mode
• goal_raw / goal_normalized: user goal and normalized intent
• selection_reasons: explainable reasons for top/bottom/shoes selection

Quick Start (Local Build)

docker build -t aiwardrobe:local .
docker run -d --name ai_wardrobe -p 8000:8000 \
  -v $(pwd)/backend/uploads:/app/backend/uploads \
  -v $(pwd)/backend/data:/app/backend/data \
  aiwardrobe:local

Using Prebuilt Image

docker pull ghcr.io/leoz9/aiwardrobe:latest
docker run -d --name ai_wardrobe -p 8000:8000 \
  -v $(pwd)/backend/uploads:/app/backend/uploads \
  -v $(pwd)/backend/data:/app/backend/data \
  ghcr.io/leoz9/aiwardrobe:latest

Docker Compose

git clone https://github.com/leoz9/AIWardrobe.git && cd AIWardrobe
docker compose up --build -d

Access at http://localhost:8000  |  API docs at http://localhost:8000/docs

Data is persisted in backend/data and backend/uploads. A few sample clothing images are included under backend/uploads/demo for quick体验.

⭐ Star History

🤝 Contributing

Contributions are welcome! Feel free to open an Issue or submit a Pull Request.

📄 License

MIT © 2024 leoz9