URL Phishing Checker

A lightweight, web-based tool that detects whether a URL is likely phishing or safe using a hybrid approach of Machine Learning (Random Forest/SVM) and heuristic rule-based analysis.

Built for educational purposes and experimentation in cybersecurity and applied AI.

Overview

Phishing attacks often rely on visual deception and URL manipulation. This tool analyzes the structure and content of a URL to determine its malicious intent.

Key Features

Real-time Analysis: Accepts a URL and processes it instantly.
Hybrid Detection: Combines ML probability with hard-coded security rules.
Explainability: Returns a confidence score and human-readable reasons (e.g., "Suspicious keywords found").
Performance: Caches results locally to reduce redundant processing.
API Support: Includes a JSON endpoint for programmatic access.

How It Works

Input: User submits a URL via the UI or API.
Normalization: The URL is cleaned and standard components are parsed.
Feature Extraction: The system extracts numerical and categorical features (e.g., length, special char count, domain age).
Prediction: * The ML Model calculates a probability score.
- Rule-based logic scans for known red flags.
Output: A JSON response is returned with the verdict and cached for future lookups.

Alt text

Setup

Clone the repo
Install deps: pip install -r requirements.txt
Download the dataset from Kaggle and place dataset_phishing.csv in the root
Run python retrain.py to generate model.pkl
Run python app.py

Tech Stack

Core: Python 3.x
Web Framework: Flask
Machine Learning: scikit-learn, pandas, joblib
Architecture: REST API + Simple HTML/CSS Frontend

⚠️ Disclaimer

Educational Use Only: This tool does not guarantee 100% protection against phishing attacks.
Predictions are probabilistic and based on the training data used.
False positives (safe sites marked as phishing) and false negatives (phishing sites marked as safe) may occur.
Do not rely on this tool as your sole security measure.

URL_CHECKER

Most Popular