ARYPHISH_DETECTOR

Made by Aryan Giri

An intelligent web tool that uses the power of Google Gemini and OpenAI's ChatGPT to analyze URLs and detect phishing risks in real-time.

---

🛑 The Problem: The "Human Firewall" is Fallible

Even well-trained employees who use VPNs, 2FA, and recognize basic attacks can still fall for social engineering. Attackers know that tricking a human is often easier than bypassing technical defenses. This is why phishing remains the #1 attack vector in real-world breaches.

---

🎯 What is Phishing?

Phishing is when attackers impersonate trusted entities to steal sensitive information such as:

• Login credentials
• Credit card information
• Personally identifiable data
• Corporate secrets

Advanced Types of Phishing

• Spear Phishing: Targeted attacks using personal info to build trust.
• Whaling: Targeting executives and administrators.
• Vishing: Voice phishing through calls.
• Smishing: Phishing through SMS messages.
• Quishing: QR code phishing.

---

🛡️ The Solution: ARYPHISH_DETECTOR

Most URL filters rely on outdated databases. Human training helps, but it isn't perfect.

ARYPHISH_DETECTOR acts as an AI-powered second opinion, analyzing:

• Deceptive forms
• Fake login pages
• Obfuscated JavaScript
• Typosquatted domains
• Threatening or urgent language

It uses multiple LLMs to inspect a website in real time — a huge advantage over static tools.

---

📸 Screenshots

---

🚀 Features

• Dual AI Mode: Analyze URLs with both Gemini and ChatGPT.
• Single AI Mode: Choose your model.
• Real-Time Content Fetching: Fetches live HTML for deep inspection.
• Clear Verdicts: Simple "Safe" or "Phishing" output.
• AI Reasoning: Understand why the site is flagged.
• Web-Based UI: Clean interface built with Flask + Tailwind.
• Stylish CLI: pyfiglet banner on startup.

---

🛠️ How It Works

1. User enters a URL and chooses an AI model.
2. Frontend sends request to Flask backend.
3. Backend uses httpx to fetch live HTML.
4. AI receives structured prompt containing URL + source code.
5. AI returns JSON response with Verdict + Reasoning.
6. Frontend displays results using cards for Gemini and ChatGPT.

---

⚙️ Installation & Setup

1. Clone the Repository

git clone https://github.com/giriaryan694-a11y/ARYPHISH_DETECTOR.git
cd ARYPHISH-DETECTOR

2. Create and Activate Virtual Environment

Linux/macOS:

python -m venv venv
source venv/bin/activate

Windows:

python -m venv venv
.\venv\Scripts\activate

3. Install Dependencies

pip install -r requirements.txt

4. Add API Keys

Create a file named keys.txt:

GEMINI_API=YOUR_GEMINI_API_KEY_HERE
CHATGPT_API=YOUR_OPENAI_API_KEY_HERE

At least one API key is required.

5. Run the Server

python main.py

The banner appears, and the server starts.

---

🖥️ Usage

1. Open browser → http://127.0.0.1:5000
2. Enter a URL.
3. Choose AI model.
4. Click Analyze.
5. View results.

---

💻 Tech Stack

• Backend: Python, Flask, httpx
• AI Models: Google Gemini, OpenAI ChatGPT
• Frontend: HTML, Tailwind CSS, Vanilla JS
• CLI: pyfiglet, colorama

---

📄 License

This project is licensed under the Apache License 2.0. See the LICENSE file for full license text.