Back to Projects

Phisher

ML-based phishing detection system with a browser extension, FastAPI backend, research workflow, and IEEE ICAC 2025 paper.

Phisher

Phisher is a dual-architecture phishing detection system that combines a browser extension with a web application to detect and analyze phishing URLs in real-time. The platform features a lightweight Chrome extension with a Random Forest classifier for immediate URL scanning, displaying warnings and threat scores when potential phishing sites are detected with customizable notification levels. The web application provides comprehensive URL analysis that is also with a Random Forest classifier, including detailed threat assessments, identified risk factors, and educational content about phishing protection strategies. The system includes an admin dashboard for monitoring phishing detections, viewing ML model performance metrics, and accessing system logs. Built with React.js and FastAPI, the architecture utilizes Docker containerization, PostgreSQL with Prisma ORM for data persistence, and scikit-learn machine learning models for both real-time extension scanning and deep URL analysis.

Research Paper

Phisher: A Dual-Architecture Phishing Detection System Using Browser Extension and Machine Learning

2025 7th International Conference on Advancements in Computing (ICAC)· 2025 · IEEE

View on IEEE

Tech Stack

Backend

  • Python
  • FastAPI
  • Express
  • Node.js

ML/AI

  • scikit-learn

Frontend

  • React
  • JavaScript
  • MUI
  • Chrome Extension

Database

  • PostgreSQL
  • Prisma

DevOps

  • Docker
  • Google Cloud Platform