Web-Based Deep Learning Model for Zero Day Vulnerability Detection using FastAPI

In the evolving world of technology, the growing number of security weaknesses is becoming a serious concern, for our online safety. This research presents a method leveraging LSTM technology to identify both familiar and undiscovered vulnerabilities in log files of web application frameworks. The platform features a user interface developed using ReactJS with upcoming upgrades focused on enhancing appeal and simplifying the log file upload process, for vulnerability assessment. By enabling users to effortlessly upload logs the system creates reports that outline forms of attacks offering valuable insights, into possible security risks. Recent live testing of the system showed its ability to quickly identify threats showcasing its potential as a proactive tool for enhancing cybersecurity defenses. Moreover, the model's prediction accuracy stands at a 98% proving its strength and dependability in identifying weaknesses. These results underscore the significance of security measures, in dealing with the evolving landscape of digital threats.