Short Summary

This project uses zero-knowledge machine learning (zkML) with EZKL to enable privacy-preserving classification of clothing designs for export tax purposes. Instead of sharing product photos with regulators, manufacturers can locally classify a design using a public ML model and generate a zero-knowledge proof of the result. The proof is submitted to tax authorities, who verify it without ever accessing the original design data. This protects intellectual property, prevents counterfeiting, ensure.

Project Purpose

The purpose of this project is to develop a privacy-preserving system for classifying clothing designs using zero-knowledge machine learning (zkML) with EZKL. This solution allows manufacturers to prove the correct tax category of their products to regulators without revealing sensitive design data. By protecting intellectual property and preventing counterfeiting, the project supports secure, efficient, and compliant international trade.

Additional Comments

See it live at https://fashion.document.legal We initially worked on a github repo but had issue with deployment, so we eventually moved to gitlab. The backend is a serverless container that may take some time to warm up.