VOICE-ASSISTED SHOPPING APPLICATION

Abstract—Voice-assisted e-commerce applications play a vitalrole in enhancing digital accessibility for visually impairedusers by enabling an inclusive, screen-free shoppingexperience. This paper presents a voice-assisted shoppingapplication that supports end-to-end e-commerce activities,including product discovery, cart management, and orderplacement through natural voice interaction. The systemfunctions as a virtual shopping assistant, capturing userintent using Speech-to-Text (STT) technology and providing backend-driven sequential product narration to deliver real-time product descriptions, pricing, and availability details. Interactive audio feedback is generated through Text-to-Speech (TTS), maintaining a continuous conversational loop with context-aware prompts such as product confirmation,cart updates, item removal, and checkout approval. Toenhance user engagement and flexibility, the applicationintegrates a Large Language Model (LLM)-based AI modulethat generates human-like, conversational responses togeneral shopping queries beyond transactional commands.Developed using Flutter for the frontend, Node.js for backendservices, and MongoDB for data storage, with AI capabilitiesenabled through OpenAI APIs, the proposed solutionimproves usability, promotes digital independence, anddemonstrates the effectiveness of voice-driven AI in reducingaccessibility barriers in online commerce for visuallyimpaired communities.