Web application security is increasingly important, particularly given the rise of phishing and spam attacks. URLs, while used to access legitimate resources, can also be exploited by attackers to compromise users accessing those sites. This project aims to safeguard users by detecting malicious websites, thereby reducing the risk of cyberattacks. Our machine learning algorithm helps prevent users from clicking malicious links by identifying potentially harmful URLs and cautioning them about the associated risks, such as threats or sensitive information leaks. Using various classification algorithms, including decision trees, random forests, logistic regression, SVM, and KNN, we discovered that the Extreme Boosting technique performed the best. The web application built using our algorithm can be beneficial in various settings to address concerns related to phishing and online security by allowing users to quickly verify URLs before interacting with potentially harmful websites. This study is essential for advancing AI-related cybersecurity research and offering practical applications of these concepts to this rapidly growing field.