PasswordStrengthML is a Python desktop application built with PyQt6 and scikit-learn that predicts the strength of passwords in real-time. The app analyzes passwords based on length, character diversity, entropy, and common patterns, providing a score from 0 to 100.
Screenshots
- Real-time password strength prediction
- Displays detailed extracted features
- Highlights common weak passwords
- Option to show/hide typed password
- Save and load trained models
- Modern gradient-based UI using QSS
- Threaded predictions to avoid UI blocking
- Clone the repository:
git clone https://github.com/Luka12-dev/PasswordStrengthML.git
cd PasswordStrengthML
Install dependencies:
bash
pip install PyQt6 scikit-learn pandas numpy joblib
Place an icon file named AI5.ico in the project directory (optional).
Usage
Run the application:
bash
python PasswordStrengthML.py
Type a password in the input field.
The app updates a strength score (0-100) in real-time.
View detailed feature breakdown including length, entropy, character counts, and pattern checks.
Optionally, check "Show password" to reveal the typed text.
Save or load trained models using the buttons provided.
Note: Do not test real passwords to ensure privacy and security.
How It Works
Feature Extraction:
Password length
Shannon entropy
Character classes: lowercase, uppercase, digits, special characters
Common pattern detection (e.g., 1234, abcd, common passwords)
Character diversity
Model:
Uses a RandomForestRegressor trained on synthetic password data
Score ranges from 0 (weak) to 100 (strong)
Threaded predictions ensure smooth UI updates
UI:
Built with PyQt6
Gradient background, modern styling via QSS
Progress bar and label show strength score
Text box shows extracted features
Development Notes
Model Training: Synthetic data is generated with common passwords and scored using entropy, diversity, and pattern checks.
Threading: Predictions run in a separate QThread to prevent GUI freezing.
Debounce: Input is debounced to reduce redundant predictions on fast typing.
Contributing
Contributions are welcome! You can:
Improve password scoring algorithms
Add more advanced password pattern detection
Enhance UI styling and interactivity
Add support for exporting results or batch analysis
License
MIT License
Disclaimer
This application is for educational and testing purposes only. Avoid entering real passwords. The generated scores are estimations and do not guarantee password security.