AI VOGUE: A Multi-Modal Textile E-Platform (MMT-EP) Integrating Generative AI, AR Visualization, and P2P Commerce

Abstract

The current e-commerce landscape is highly fragmented, forcing users to switch between multiple platforms to meet different needs such as retail shopping, apparel rentals, and custom design services. This paper introduces AI VOGUE, a unified Multi-Modal Textile E-Platform (MMT-EP) designed to eliminate these inefficiencies through a cohesive Systemic Integration Architecture. Unlike conventional solutions that offer isolated features, AI VOGUE seamlessly integrates Generative AI–driven Virtual Try-On (VTO), AR-based accessory visualization, LLM-powered conversational styling, and a Multi-Source Aggregation Engine into one streamlined experience. The platform employs a hybrid technology stack, using Three.js for high-fidelity 2D and AR rendering, MediaPipe and Google ARCore for real-time accessory tracking, the Groq API for ultra-low-latency intelligent responses, and Appwrite as a scalable serverless backend. A major contribution of this work is the development of a hybrid VTO pipeline that combines the Gemini API with the Nano Banana Model to produce photorealistic garment transfer results, along with the creation of a decentralized “Creative Threads” marketplace that supports peer-to-peer designer interactions. Overall, this research shows that integrating diverse computational models within a unified, user-centric system significantly enhances user engagement while democratizing access to advanced fashion technologies.

KEYWORDS:- Multi-Modal E-Platform, Generative AI, Virtual Try-On, Augmented Reality, Serverless Architecture, Conversa- tional Commerce.