Leveraging Gemini API for Cyber Threat Detection in Emails

This project develops an Email Analysis and Phishing Detection System with distinct Admin and User roles to classify emails, assess risks, and detect phishing attempts. Leveraging Google’s Gemini AI, the system performs advanced email content analysis and threat classification, ensuring robust security against email-based threats. The Admin oversees the platform via a comprehensive dashboard, managing users, reviewing submissions, and accessing statistical reports for system health and performance monitoring. Users submit emails for analysis, view results, receive actionable guidance, and provide feedback to enhance the platform. The system categorizes emails as personal, promotional, transactional, phishing, spam, or business, using AI-driven natural language processing and regular expressions to identify suspicious keywords, URLs, and sender domain reputation. Threat levels (High, Medium, Low, Unknown) are assigned based on these analyses, with phishing detection focusing on content, headers, and malicious indicators. The frontend, built with React, Tailwind CSS, and Font Awesome, ensures a dynamic, responsive user interface, while the backend, powered by Node.js, Express, and MySQL, supports efficient API routing and data management. The integration of Gemini AI enables precise sentiment and intent analysis, enhancing classification accuracy. This scalable system offers a proactive solution for real-time email threat detection, with future enhancements planned to incorporate advanced machine learning and real-time threat intelligence to combat evolving email-based threats effectively.

Keywords: Email Analysis, Phishing Detection, Generative AI, Email Classification, MERN