Authors: Mr. Karthiban R, Dhayalan K, Akshita K, Jerisha Flavio J, Kalaiselvi S
Abstract: As digital learning environments continue to evolve, maintaining secure and focused internet usage has become a critical requirement for institutions and organizations. Existing browser monitoring tools often lack real-time visibility and are unable to detect VPN-based evasion techniques, which users exploit to bypass access restrictions. To address these limitations, this work proposes an intelligent browser activity monitoring and VPN detection system featuring a centralized administrative dashboard. Built on a Flask-based backend, the system securely gathers and visualizes browsing data through interactive charts and tables. A machine learning model continuously refines detection by learning administrative preferences—distinguishing between authorized and unauthorized sites—and improving decision accuracy over time. The adaptive framework enhances detection precision by integrating AI-driven behaviour learning with network anomaly analysis. By evaluating parameters such as IP consistency, latency fluctuations, and metadata patterns, the system effectively identifies tunnelling or masked connections even in encrypted networks. Its modular and cross-platform architecture ensures seamless data flow between clients and the central dashboard while preserving privacy and performance. Designed for scalability and reliability, the solution provides administrators with actionable insights and real-time control, making it an effective tool for maintaining policy compliance and secure browser activity in educational and institutional environments.
DOI: http://doi.org/10.5281/zenodo.17529567
