Real-Time Malicious URL and Inappropriate Content Blocking System

In today's digital age, malicious URLs present a persistent threat to cybersecurity, enabling phishing, malware distribution, and data breaches.This paper introduces a comprehensive system that combines advanced machine learning algorithms with natural language processing (NLP) to detect and classify malicious URLs accurately.The system backend is developed using Python Flask, while the frontend leverages Node.js, with PostgreSQL managing the database.It integrates external threat intelligence APIs such as Google Safe Browsing and VirusTotal to augment detection reliability.Core machine learning models-Random Forest, Principal Component Analysis (PCA), and Naive Bayes-are employed for feature extraction and classification, while a child NLP model provides context-aware analysis of URL content, controllable exclusively by administrators to maintain operational security.The detection model is initially trained on 120,000 URL samples and later scaled to a dataset of 300,000 with an achieved accuracy of 98.2%.The system supports automated daily retraining to accommodate new threat patterns and dataset updates, ensuring adaptability in an evolving cyber threat landscape.Practical deployment is exemplified through a Chrome extension, delivering real-time URL safety assessment to end-users.This admin-controlled, scalable system offers a robust, efficient, and secure solution for mitigating web-based cyber threats.