International Journal of Science, Engineering and TechnologyAuthors: Nareandra Sanjay Bhute, Rohan Bhanudas Pawar, Rutik Nandkishor Pawar, Aditya Madan Mapari, Prof. N. N. Kumbhar
Abstract: The rapid proliferation of online travel services has resulted in a fragmented ecosystem where travelers must navigate multiple disconnected platforms to book accommodation, rent vehicles, and discover dining options. This paper presents the design and implementation of WanderLust, a full-stack, multi-category travel booking platform that consolidates stays, vehicle rentals, and dhaba (local dining) reservations into a single unified web application. The system is built on a Node.js v18 and Express.js v4 backend with MongoDB Atlas as the NoSQL database, employing a Model-View-Controller (MVC) architecture with EJS server-side rendering. Key implemented features include: (i) an AI-powered trip planning module and conversational chatbot leveraging the OpenAI GPT-4o-mini API with function calling for real-time database queries; (ii) dual payment gateway integration with Stripe (international) and Razorpay (Indian UPI, net banking) alongside an in-app digital wallet; (iii) real-time host–guest communication via Socket.IO WebSockets; (iv) geospatial proximity search using Mapbox SDK with MongoDB 2dsphere indexing; (v) a comprehensive security layer comprising Helmet.js Content Security Policy, CSRF token validation, rate limiting, XSS sanitization, and NoSQL injection prevention; (vi) a multi-channel notification system spanning email (Nodemailer/Brevo API), browser push notifications (Web Push), and in-app alerts; and (vii) cross-platform mobile deployment via Capacitor for native Android packaging. The platform manages 14 Mongoose data models, 24 route modules, and 20 utility services. Authentication is handled through Passport.js with local credentials and Google OAuth 2.0, augmented by OTP-based email verification. Automated testing was conducted using Jest and Supertest with an in-memory MongoDB instance. The application is deployed on Render (PaaS) with Cloudinary CDN for media delivery and Sentry for real-time error monitoring. Evaluation demonstrates a feature-complete, production-grade platform that addresses the identified research gap of no existing unified open-source solution integrating multi-category bookings, AI-assisted itinerary generation, dual-gateway Indian payment support, and hybrid mobile deployment within a single cohesive architecture.
DOI:
