About the project
Wardrobe is a private web application for digitizing, organizing and using a personal clothing collection. The project is based on the open-source tandpfun/wardrobe application, which I significantly adapted, extended and prepared for secure production use on my own infrastructure.
The application allows users to upload clothing photographs and automatically detect individual garments using the OpenAI Responses API. After detection, the OpenAI Images API is used to generate clean product cutouts. The system supports uploading up to 20 images at once, including HEIC and HEIF photographs from iPhones, with a visible processing queue and manual result review before saving.
Users can browse and search their digital wardrobe and edit garment names, categories, subcategories, colours, brands and other item information. The system supports manually building outfits and locally generates suggestions based on garment categories and colour compatibility without requiring an additional API request.
The application also suggests outfits based on weather conditions and a described occasion. An AI virtual try-on can be generated for an individual garment or a complete outfit. Image generation is performed only after explicit user confirmation.
As part of my adaptation, I redesigned the React user interface and implemented an Express production server, MySQL persistence, private image storage, database migrations, backups and production deployment support for Hostinger.
Special attention was given to security and privacy. The system includes server-side sessions, secure HttpOnly and SameSite cookies, CSRF protection, login rate limiting, security headers, uploaded image validation, metadata removal, protected API routes and private delivery of all user photographs.
The project includes detailed technical documentation covering architecture, data flow, security, privacy, database design, storage, testing, Hostinger deployment, backup and restoration procedures.
Key features include AI garment detection, batch photo upload, HEIC/HEIF processing, cutout generation, outfit suggestions and virtual try-on.