A Real-Time Full-Stack Chat Application

Abstract

The rapid evolution of web technologies has fun- damentally shifted user expectations regarding digital commu- nication. Modern web users demand instantaneous, low-latency messaging platforms that operate seamlessly without the need for manual page refreshes. Traditional HTTP request-response models, relying on long-polling or short-polling, often suffer from high server overhead, significant latency, and inefficient resource utilization when scaled for real-time chat environments. This research presents the design, development, and implementation of a highly scalable, real-time Full-Stack Chat Application leveraging the MERN stack (MongoDB, Express.js, React.js, and Node.js) integrated with Socket.io. The proposed platform provides instantaneous bidirectional communication, distinct global state management for authenticated sessions, live online/offline user status tracking, and secure media sharing capabilities. A critical architectural advancement in this system is the implementation of an event-driven WebSocket layer that maps active socket connections to unique user identities, ensuring precise message broadcasting and real-time UI synchronization. Developed utilizing Vite, Tailwind CSS for a mobile-first responsive layout, JSON Web Tokens (JWT) for secure au- thentication, and Cloudinary for global asset management, the system guarantees a fluid user experience and robust backend persistence. Experimental analysis demonstrates that the event- driven architecture drastically reduces message delivery latency compared to traditional HTTP polling, ensures reliable data persistence in MongoDB, and provides an optimized environment for seamless digital communication.