Train your team before attackers do.

Python 3.11 Django 5 React 19 PyTorch LSTM EN + AR (RTL)
๐Ÿ“ฆ

The live demo has been retired

PhishAware was a senior capstone project. The hosted instance has been shut down, but the full source code, architecture, and documentation remain available on GitHub for anyone interested in how it was built.

View on GitHub Learn more โ†“

What PhishAware did

๐ŸŽฏ

Phishing Simulations

Real emails with tracking pixels and lure links. Teaches at the moment of failure with a "caught" page that explains the red flags.

๐Ÿค–

AI Email Generator

Two custom-trained LSTM models (English + Arabic) built from scratch in PyTorch. No third-party LLM wrapper.

๐Ÿ“Š

Live Risk Scoring

Per-employee risk score recalculated the moment a tracking event fires. Composite of simulations, quizzes, and training.

๐Ÿ“š

Adaptive Training

Modules auto-assigned based on each employee's risk score โ€” high-risk users get more targeted remediation.

๐Ÿ†

Gamification

Badges, points, and leaderboards drive engagement. Because nobody finishes boring security training.

๐ŸŒ

Fully Bilingual

English and Arabic with proper RTL layout โ€” designed in from day one, not bolted on as an afterthought.

Tech Stack

Backend

  • Python 3.11
  • Django 5 + DRF
  • SimpleJWT
  • SendGrid SMTP
  • SQLite โ†’ PostgreSQL

Frontend

  • React 19
  • Vite 7
  • Tailwind 3
  • React Router 6
  • i18next + RTL

AI / Data

  • PyTorch (custom LSTM)
  • 14k EN samples
  • 10k AR samples
  • Bilingual tokenizer
  • Trained from scratch

The Team

Emad Alzahrani
Talal Alharbi
Thamer Alkhatheri

Senior capstone โ€” built because the bar should be higher than "it works on localhost."