Sora2 Watermark Cleaner (GUI) is a graphical application for removing your own technical or test watermarks from images and videos during development or internal workflows.
- Features
- Requirements
- Installation
- Launching the GUI
- GUI Overview
- Advanced Settings
- Batch Processing
- GPU Acceleration
- Project Structure
- Development & Testing
- Troubleshooting
- Contributing
- License
- Example main.py (GUI launcher)
- Fully graphical interface (PyQt / Tkinter depending on build)
- Drag & drop images or videos
- Batch folder support
- Real-time preview (before/after)
- Adjustable cleaning strength
- Fast/Standard/Deep modes
- GPU acceleration option
- Progress bar, logs and error reporting
- Saves settings between sessions
Python 3.9+ Recommended: GPU with CUDA (optional)
git clone https://github.com/ossessd/SoraWatermarkCleaner.git cd SoraWatermarkCleaner pip install -r requirements.txt
python main.py
or if using a packaged build:
./Sora2WatermarkCleaner (Linux) Sora2WatermarkCleaner.exe (Windows) Sora2WatermarkCleaner.app (macOS)
Main window contains:
-
Input section
- Drag & drop area
- “Browse File” and “Browse Folder”
- Input preview
-
Output section
- Output path selector
- Auto-naming mode
-
Processing settings
- Strength (slider 0–100%)
- Mode: Fast • Standard • Deep
- GPU enable switch
- Preserve folder structure (batch)
-
Real-time preview
- Before / After toggle
- Auto-refresh
-
Action buttons
- “Clean file”
- “Clean folder”
- “Stop”
-
Status & logs
- Progress bar
- Realtime logs panel
- Error console
In Settings → Advanced:
• Parallel workers count
• Temporary cache directory
• Preview resolution limit
• Auto-save results
• Auto-open output folder
• Detailed logging
Configuration is stored in:
Windows: %APPDATA%/Sora2WC/config.json
Linux: ~/.config/Sora2WC/config.json
macOS: ~/Library/Application Support/Sora2WC/config.json
Open the folder in the GUI:
- Click “Select Folder”
- Choose processing mode
- Enable options:
• Recursive
• Preserve file structure
• Skip unreadable files - Click “Start Batch”
The progress bar shows:
- current file
- overall progress
- success/skipped/error count
Enable by switching “GPU Mode” in the GUI (if supported).
Deep mode is significantly faster with GPU.
If GPU is not detected:
- Check CUDA installation
- Update GPU drivers
- Restart the application
pip install -r requirements-dev.txt pytest -q
Linters: black src gui tests flake8 src gui tests
Problem: Preview is slow
Fix: Lower preview resolution in Settings
Problem: GPU not detected
Fix: Reinstall CUDA, enable GPU in settings
Problem: Output is distorted
Fix: Reduce strength or switch to Standard mode
Problem: Video fails to export
Fix: Install ffmpeg
- Fork the repository
- Create feature branch
- Add tests
- Submit PR
MIT (or your chosen license)