In today’s e-commerce landscape, customers often face uncertainty about how clothes will look on them, leading to dissatisfaction and high product return rates. CustomCart is an AI-powered personalized e-commerce platform designed to address this challenge by integrating a virtual try-on system. Using generative AI (via COMFYUI), the system allows users to upload their own photos and see dresses realistically applied to them, simulating a real fitting experience.

The platform combines:

Generative AI for virtual try-on and personalized product suggestions

A lightweight web front-end storing cart data in localStorage

Multiple backend services (Node.js, FastAPI, and COMFYUI) working together seamlessly

By providing a personalized and visual shopping experience, CustomCart aims to reduce return rates, improve customer satisfaction, and make online shopping more engaging and interactive. Also used a sample RazarPay gateway using Razorpay API.

The COMFYUI workflow used for clothes swapping is given bellow:

There is a custom workflow made for each apparel, which needs to be changed in future works, so that any apperal can be given as an input using a simple websocket.