PhishGuard AI

Advanced ML-Powered Phishing Detection System

0
Total Scans
0
Threats Blocked
95.2%
ML Accuracy
3.2s
Avg Scan Time
Guest

Scan a Website URL

Quick Test:
🛡️

PhishGuard AI — Complete Guide

Your comprehensive phishing protection assistant

📋 How to Use PhishGuard AI

1

Enter URL

Copy and paste any suspicious URL into the input field. Make sure it starts with http:// or https://

💡 Tip: Use the quick test buttons for instant examples!
2

Click Scan

Press the "Scan Website" button or use keyboard shortcut Ctrl + Enter

⚡ Speed: Analysis typically completes in under 1 second!
3

Review Results

Check the detailed analysis including:

  • Classification: Phishing / Suspicious / Safe
  • Confidence Score: 0–100% accuracy
  • Risk Level: Critical / High / Medium / Low
  • Detection Breakdown: Multi-module analysis
💡 Check the risk factors section for detailed information
4

View Detailed Analysis

Click "View Detailed Analysis" to see:

  • 📊 Multi-Module Detection Chart: 6 detection methods visualised
  • 🎯 Ensemble Breakdown: How each module contributed
  • 📈 URL Feature Analysis: Feature-level bar chart
  • ⚠️ Risk Factors: Specific security concerns found
  • 🔎 Metrics: Domain age, HTTPS status, URL length

⌨️ Keyboard Shortcuts

Scan URL Ctrl + Enter
Close Guide / Sidebar Esc

🔄 System Flow

📥

1. URL Input

User submits URL for analysis

↓
✅

2. URL Validation

Pre-flight checks: Syntax, DNS, Domain existence

↓
🧠

3. ML Model Analysis

Random Forest classifier (60% weight)

  • 11 URL features extracted
  • Trained on 10,000+ samples
  • 95%+ accuracy rate
↓
🔍

4. Lexical Analysis

URL structure & patterns (15% weight)

  • Entropy calculation
  • Homoglyph detection
  • Suspicious TLD check
↓
🌐

5. Domain Reputation

Domain verification (15% weight)

  • DNS resolution check
  • Whitelist verification
  • Brand impersonation detection
↓
📄

6. Behavior Analysis

Webpage patterns (5% weight)

  • Suspicious path detection
  • Query parameter analysis
  • Obfuscation check
↓
💬

7. NLP Analysis

Phishing language detection (5% weight)

  • Urgency keywords
  • Trust exploitation terms
  • Social engineering patterns
↓
🎯

8. Ensemble Decision

Weighted average of all 6 modules

Final Score = (ML × 0.6) + (Lexical × 0.15) + (Reputation × 0.15) + (Behavior × 0.05) + (NLP × 0.05)
• Phishing: ML Score ≥ 75%
• Suspicious: 40% ≤ ML Score < 75%
• Safe: ML Score < 40%
↓
📊

9. Result Display

Comprehensive analysis with visual charts

🔒 Security Note: All analysis happens server-side. Your URLs are not stored or shared.

Note: While we display analysis from multiple modules (Lexical, Reputation, Behavior, NLP) for transparency, the final verdict is driven entirely by the ML model score.

🛡️ About PhishGuard AI

What is PhishGuard AI?

PhishGuard AI is an advanced, AI-powered phishing detection system that protects users from malicious websites using cutting-edge machine learning and ensemble detection techniques.

🎯 Key Features

🧠

Ensemble Detection

6 detection modules working together for maximum accuracy

⚡

Real-Time Analysis

Instant results in under 1 second

📊

Explainable AI

Detailed breakdown of why a site is flagged

🔒

Privacy First

No URL logging or data collection

📈

95%+ Accuracy

Trained on 10,000+ real phishing samples

🌐

Browser Extension

Auto-scan protection as you browse

🔬 Technology Stack

Backend: Flask, Python 3.8+
ML Model: Random Forest (scikit-learn)
Frontend: HTML5, CSS3, Vanilla JavaScript, Chart.js 4.0+
Extension: Chrome Manifest V3
Database: SQLite with bcrypt & JWT

📈 Performance Metrics

95.2%
Accuracy
94.8%
Precision
96.1%
Recall
<1s
Scan Time

⚠️ Important Notes

  • Always verify suspicious links before clicking
  • Enable the browser extension for continuous protection
  • Report false positives to help improve the system
  • This tool is a detection aid, not a 100% guarantee
  • Use common sense alongside AI detection

📦 Version Information

Version: 7.0 — Unified Edition

Release Date: February 2026

Model Version: Random Forest v1.0

License: MIT