SafeNet: AI-Powered Browser Extension for Real-Time Phishing Site Detection, user Protection, and Instant Warning Alerts

Of particular concern in terms of cyber attack is phishing and other hackers because they are busy adapting URLs, domain configuration and deceptive content patterns, as a means of making it off-the-shelf blacklists and rule sets. The paper presents a phishing detection system (SafeNet) built on AI that works by examining URLs on-the-fly and can issue a warning to the computer user. It involves a combination of URL and domain-oriented features record with the aided machine learning models, including Neural Network, K-Nearest Neighbors, Naive Bayes, random forest and Support Vector Machine. The preprocessing, as well as the choice of the model is also performed with the help of Principal Component Analysis and Recursive Feature Elimination to enhance the quality of the features and remove redundancy. Its back-end prediction is based on FastAPI platform and it has a real time front-end interface based on URL submission and classifying. With experimental comparison of multiple classifiers, the model of the Neural Network displays the highest overall performances as compared to the methods being tested and is more accurate, more precise, recalls more, and F1- score better. According to the study, the main value of DanceNet is not only that it integrates multi-model phishing detection, feature optimization, and deployable real-time prediction in a single workflow, but also offers this workflow as a service. The findings show that the suggested framework can be used to facilitate the feasible phishing detection as well as to offer an opportunity to expand it in the future via ongoing feedback based enhancement.