Inspiration
The fashion industry produces millions of tonnes of textile waste every year, and a major contributor is overconsumption. Many people own clothing items they rarely wear simply because they forget what they already have.
We were inspired to create ClosetNinja to help people build a more mindful relationship with their wardrobe — encouraging them to use, reuse, repair, and appreciate the clothes they already own before buying more.
What We Built
ClosetNinja is an AI-powered digital wardrobe that transforms a user's physical wardrobe into an intelligent, searchable collection.
Users can upload images of their clothing, and our computer vision system identifies important attributes such as:
- Clothing type (shirt, jacket, jeans, sweater, etc.)
- Colour
- Style
- Season suitability
- Usage history
The application also allows users to log their Outfit of the Day (OOTD). By analysing outfit images, ClosetNinja tracks which items are worn over time and builds insights into clothing habits.
When clothing items have not been used for extended periods, ClosetNinja encourages sustainable actions:
- Generate outfit suggestions to bring forgotten items back into rotation
- Donate or sell unused clothing
- Repair or repurpose damaged clothing
We also developed a mindful purchasing feature that helps users evaluate new purchases by converting price into working hours and identifying whether they already own similar items.
How We Built It
ClosetNinja combines modern web technologies with AI-powered image understanding to create an intelligent digital wardrobe.
For clothing recognition, we initially experimented with YOLO-based object detection models. However, achieving reliable clothing attribute extraction — such as clothing type, colour, and style — required more training data and fine-tuning than was possible within the hackathon timeframe.
We then integrated the Claude API with vision capabilities, allowing us to perform flexible image understanding and extract meaningful clothing attributes from uploaded images. This helped us move beyond simple object detection and build a richer wardrobe intelligence system.
Frontend
- Next.js
- React
- TypeScript
- Tailwind CSS
- React Query for efficient data fetching and caching
Backend
- FastAPI
- Python
- Claude API for vision-based clothing analysis
- Supabase for authentication and storage
Key Features Implemented
- User authentication
- Digital wardrobe creation through image uploads
- AI-powered clothing analysis
- Outfit of the Day (OOTD) tracking
- Wardrobe history and usage insights
- Sustainable decision-making tools
Challenges We Faced
One of our biggest challenges was building a reliable computer vision pipeline while maintaining a simple user experience.
Initially, we explored YOLO for clothing detection, but adapting it to accurately understand fashion-specific attributes such as style, season, and colour required extensive model training.
We overcame this by pivoting to Claude's vision capabilities, allowing us to focus on building a complete sustainability-focused experience rather than only solving the classification problem.
Another challenge was designing a system that goes beyond identifying clothes. We wanted ClosetNinja to understand user behaviour and encourage sustainable decisions, rather than simply acting as an image classifier.
Integrating frontend, backend, authentication, AI services, and storage together within a limited hackathon timeframe was another key engineering challenge.
What We Learned
Through building ClosetNinja, we learned how AI can be applied to solve real-world sustainability problems.
We improved our skills in:
- Full-stack development
- API integration
- Computer vision workflows
- AI-powered application design
- Building user-focused technology solutions
Most importantly, we learned that technology can influence everyday decisions and help create more sustainable habits.
Future Improvements
Some features we would like to add include:
- AI outfit recommendations based on weather
- Calendar-based clothing usage tracking
- Cost-per-wear analytics
- Wardrobe diversity insights
- Similar-item detection before purchasing
- Gamification and sustainability achievements
Impact
ClosetNinja aims to reduce unnecessary clothing consumption by helping people make better use of what they already own.
Instead of asking: "What should I buy next?" ClosetNinja helps users ask: "How can I make the most of what I already have?"
Built With
- ai
- claudeapi
- fastapi
- nextjs
- python
- supabase
- typescript
Log in or sign up for Devpost to join the conversation.