SpendWise: An AI-Powered Personal and Group Expense Management System for Students

This paper presented SpendWise, an AI-powered personal and group expense management system designed for student and young professional users. The proposed system demonstrates the feasibility of integrating modern full-stack web technologies with locally deployed Large Language Models to improve financial awareness, budgeting discipline, and collaborative expense management within a unified application. The architecture combines React-based frontend visualization, FastAPI backend services, MongoDB data persistence, and Ollama-hosted LLaMA inference to deliver personalized financial analytics and group settlement functionality. Experimental evaluation confirmed stable backend performance with average API response latency of 120 ms, analytics rendering time of 340 ms, and AI insight generation averaging 2.1 s under standard workload conditions. The AI financial insight engine demonstrated effective generation of personalized savings recommendations, overspending alerts, and behavioral spending classifications using a preprocessing-optimized inference pipeline. Database benchmarking across 1,000 sequential operations confirmed low-latency performance appropriate for real-world student-scale deployments. Future work will prioritize UPI and payment gateway integration, predictive financial forecasting using historical expenditure data, OCR-based receipt scanning for automated expense entry, cloud-hosted AI inference optimization, and advanced spending anomaly detection. Extended evaluation using longitudinal user datasets and field deployments will further validate recommendation accuracy and behavioral personalization capabilities. Overall, SpendWise contributes a replicable design framework for AI-integrated financial management systems targeting underserved student demographics through accessible, intelligent, and collaborative web-based platforms.

[1]  H. Touvron et al., “LLaMA: Open and Efficient Foundation Language Models,” arXiv preprint arXiv:2302.13971, Feb. 2023. [2]  S. Ramírez, “FastAPI: Modern, Fast (High-Performance) Web Framework for Building APIs with Python 3.7+,” FastAPI Documentation. [Online]. Available: https://fastapi.tiangolo.com/. Accessed: May 2026. [3]  MongoDB, Inc., “MongoDB Manual Documentation,” MongoDB Documentation. [Online]. Available: https://www.mongodb.com/docs/. Accessed: May 2026. [4]  Meta Open Source, “React – The Library for Web and Native User Interfaces,” React Documentation. [Online]. Available: https://react.dev/. Accessed: May 2026. [5]  Vite Team, “Vite Next Generation Frontend Tooling,” Vite Documentation. [Online]. Available: https://vitejs.dev/. Accessed: May 2026. [6]  Tailwind Labs, “Tailwind CSS: A Utility-First CSS Framework,” Tailwind CSS Documentation. [Online]. Available: https://tailwindcss.com/docs. Accessed: May 2026. [7]  Recharts Team, “Recharts – Redefined Chart Library Built with React and D3,” Recharts Documentation. [Online]. Available: https://recharts.org/. Accessed: May 2026. [8]  Framer B.V., “Framer Motion – Production-Ready Motion Library for React,” Framer Motion Documentation. [Online]. Available: https://www.framer.com/motion/. Accessed: May 2026. [9]  Ollama, “Ollama: Get Up and Running with Large Language Models Locally,” Ollama Documentation. [Online]. Available: https://ollama.com/. Accessed: May 2026. [10]  R. H. Thaler and C. R. Sunstein, Nudge: Improving Decisions About Health, Wealth, and Happiness. New Haven, CT, USA: Yale University Press, 2008.