Skip to content

hemakeshwar/PhishGuard-AI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🛡️ PhishGuard AI

An advanced Machine Learning-powered cybersecurity platform for detecting phishing, malware, defacement, and suspicious URLs through intelligent behavioral analysis and threat scoring.


🚀 Overview

PhishGuard AI is a web-based cybersecurity application designed to identify potentially dangerous URLs before users interact with them.

The system uses Machine Learning to analyze URL behavior, structure, entropy, suspicious keywords, domain characteristics, and multiple security indicators to determine whether a URL is safe or malicious.

Unlike traditional blacklist-based solutions, PhishGuard AI focuses on behavioral analysis and feature extraction, allowing it to detect previously unseen threats.


✨ Features

  • 🔍 AI-Powered URL Analysis
  • 🛡️ Phishing Detection
  • ☣️ Malware Detection
  • ⚠️ Defacement Detection
  • 📊 Risk Score Calculation
  • 🧠 Explainable Threat Analysis
  • 📈 Analytics Dashboard
  • 📚 Scan History
  • ⭐ Favorites System
  • 📄 JSON Report Export
  • 🎨 Multi-Theme Support
  • 📱 Responsive Mobile Design
  • ⚡ Real-Time Threat Assessment

🎨 Available Themes

  • Dark Theme
  • Light Theme
  • AMOLED Theme
  • Cyber Theme
  • Matrix Theme
  • Purple Theme
  • Red Theme

🧠 Machine Learning Model

Algorithm

Random Forest Classifier

Dataset Size

707,514 URLs

Threat Classes

Class| Description Safe| Legitimate URLs Phishing| Credential Theft Attempts Malware| Malicious Software Distribution Defacement| Compromised or Altered Websites

Feature Engineering

The model extracts and analyzes multiple URL features including:

  • URL Length
  • Domain Length
  • Path Length
  • Query Length
  • Subdomain Count
  • Digit Count
  • Dot Count
  • Slash Count
  • Hyphen Count
  • Underscore Count
  • Entropy Score
  • HTTPS Usage
  • Suspicious Keywords
  • IP Address Detection
  • And More...

📂 Project Structure

PhishGuard-AI/

├── app.py ├── requirements.txt │ ├── datasets/ │ └── master_urls.csv │ ├── ml/ │ ├── train_model.py │ ├── predictor.py │ ├── feature_extractor.py │ └── explanation_engine.py │ ├── models/ │ ├── static/ │ ├── css/ │ └── js/ │ ├── templates/ │ └── index.html │ ├── screenshots/ │ └── README.md


📸 Screenshots

Screenshots for every theme and major feature are available in the "screenshots/" directory.

Included:

  • Homepage
  • Safe URL Detection
  • Phishing Detection
  • Analytics Dashboard
  • Theme Showcase
  • Scan History
  • Report Export
  • Mobile View

⚙️ Installation

Clone Repository

git clone https://github.com/hemakeshwar/PhishGuard-AI.git cd PhishGuard-AI

Create Virtual Environment

python -m venv venv

Activate Environment

Linux:

source venv/bin/activate

Windows:

venv\Scripts\activate

Install Dependencies

pip install -r requirements.txt

Run Application

python app.py


🔧 Training The Model

To train the model locally:

python ml/train_model.py

Generated model files will be stored inside:

models/


📌 Important Note

The trained machine learning model files (".pkl") are not included in this repository.

The original trained model exceeded GitHub's file size limitations and therefore cannot be uploaded directly.

To generate the model locally, run:

python ml/train_model.py

This will automatically create:

models/phishing_model.pkl models/label_encoder.pkl

The complete source code, datasets, feature extraction pipeline, backend logic, and frontend implementation are included in this repository.


🎯 Future Improvements

  • Deep Learning Models
  • Browser Extension
  • Email Phishing Detection
  • QR Code Scanner
  • Real-Time Threat Intelligence Integration
  • Cloud Deployment
  • REST API Service
  • User Accounts & Authentication

👨‍💻 Author

Hemakeshwar

MCA Student | AI & Cybersecurity Enthusiast


⭐ Support

If you found this project useful, consider giving it a Star ⭐ on GitHub.

About

Advanced phishing detection platform trained on 707,514 real-world URLs, featuring AI-powered threat analysis, risk scoring, analytics, and multi-theme security dashboards.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors