README.md

metadata

title: Phishing URL Detection
emoji: 🛡️
colorFrom: red
colorTo: yellow
sdk: gradio
sdk_version: 4.44.0
app_file: app.py
pinned: false
python_version: '3.12'

Phishing URL Detection

A machine learning-powered phishing detection app that analyzes URLs using URLScan.io and predicts whether they are phishing or legitimate.

Features

• Real-time URL Analysis: Submit any URL for instant phishing detection
• URLScan.io Integration: Automatically scans URLs and extracts security features
• ML-Powered Predictions: Uses trained models from Hopsworks Model Registry
• Detailed Results: Shows confidence scores and extracted security features

How It Works

1. Enter a URL in the web interface
2. The app scans the URL using URLScan.io API
3. Extracts security features (domain age, TLS certificate, secure requests, etc.)
4. Runs inference using a trained machine learning model
5. Displays prediction with confidence score and feature analysis

Model Information

The model analyzes these security features:

• Domain Age: How old the domain is (days)
• Secure Percentage: Percentage of HTTPS requests
• Umbrella Rank: Cisco Umbrella popularity ranking
• TLS Certificate: Certificate validity period
• URL Length: Length of the URL
• Subdomain Count: Number of subdomains

Configuration

This app requires the following environment variables (configured as Space secrets):

• HOPSWORKS_API_KEY: Hopsworks API key for model loading
• HOPSWORKS_PROJECT: Hopsworks project name
• URLSCAN_API_KEY: URLScan.io API key for URL scanning

Disclaimer

This tool is for educational and research purposes only. Predictions are not 100% accurate and should not be the sole basis for security decisions.