⠀
⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣾⠳⠶⣤⡄⠀⠀⠀⠀⠀⢀⣶⣄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢰⡇⠀⠀⣸⠃⠀⠀⠀⠀⣴⠟⠁⠈⢻⣦⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣿⠀⠀⢠⡟⠀⠀⠀⢠⡾⠃⠀⠀⣰⠟⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀ ⠀⠀⠀⠉⠛⠓⠾⠁⠀⠀⣰⠟⠀⠀⢀⡾⠋⠀⠀⠀⢀⣴⣆⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣀⣀⣀⣠⣤⣤⣤⣄⣀⣀⡀⠀⠀⠀⠀⠀⠀⠀⠙⠳⣦⣴⠟⠁⠀⠀⣠⡴⠋⠀⠈⢷⣄⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣀⣤⣴⣶⣿⣿⣿⣿⡿⠿⠿⠿⠿⠿⠿⣿⣿⣿⣿⣷⣦⣤⣀⠀⠀⠀⠀⠀⠀⠀⠀⣠⡾⠋⠀⠀⢀⣴⠟⠁⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢀⣠⣶⣿⣿⡿⠟⠋⠉⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠈⠉⠙⠻⢿⣿⣿⣶⣄⡀⠀⠀⠀⠺⣏⠀⠀⣀⡴⠟⠁⢀⣀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣠⣶⣿⣿⠿⠋⠁⠀⢀⣴⡿⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢶⣬⡙⠿⣿⣿⣶⣄⠀⠀⠙⢷⡾⠋⢀⣤⠾⠋⠙⢷⡀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣠⣾⣿⡿⠋⠁⠀⠀⠀⢠⣾⡟⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⣰⣦⣠⣤⠽⣿⣦⠈⠙⢿⣿⣷⣄⠀⠀⠀⠺⣏⠁⠀⠀⣀⣼⠿⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⢠⣾⣿⡿⠋⠀⠀⠀⠀⠀⣰⣿⠟⠀⠀⠀⢠⣤⠀⠀⠀⠀⠀⠀⠀⠀⠉⠉⠉⣿⣧⠀⠀⠈⢿⣷⣄⠀⠙⢿⣿⣷⣄⠀⠀⠙⣧⡴⠟⠋⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⣴⣿⣿⠏⠀⠀⠀⠀⠀⠀⢷⣿⡟⠀⣰⡆⠀⢸⣿⠀⠀⠀⠀⠀⠀⠀⠀⣀⡀⠀⣿⣿⡀⠀⠀⠈⢿⣿⣦⠀⠀⠙⢿⣿⣦⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⣼⣿⡿⠁⠀⠦⣤⣀⠀⠀⢀⣿⣿⡇⢰⣿⠇⠀⢸⣿⡆⠀⠀⠀⠀⠀⠀⠀⣿⡇⠀⢸⣿⣿⣆⠀⠀⠈⣿⣿⣧⣠⣤⠾⢿⣿⣧⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⣸⣿⣿⣵⣿⠀⠀⠀⠉⠀⠀⣼⣿⢿⡇⣾⣿⠀⠀⣾⣿⡇⢸⠀⠀⠀⠀⠀⠀⣿⡇⠀⣼⣿⢻⣿⣦⠴⠶⢿⣿⣿⣇⠀⠀⠀⢻⣿⣧⣀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⢀⣿⣿⣿⣿⠇⠀⠀⠀⠀⠀⢠⣿⡟⡌⣼⣿⣿⠉⢁⣿⣿⣷⣿⡗⠒⠚⠛⠛⢛⣿⣯⣯⣿⣿⠀⢻⣿⣧⠀⢸⣿⣿⣿⡄⠀⠀⠀⠙⢿⣿⣷⣤⣀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⢸⣿⣿⣿⠏⠀⠀⠀⠀⠀⠀⢸⣿⡇⣼⣿⣿⣿⣶⣾⣿⣿⢿⣿⡇⠀⠀⠀⠀⢸⣿⠟⢻⣿⣿⣿⣶⣿⣿⣧⢸⣿⣿⣿⣧⠀⠀⠀⢰⣷⡈⠛⢿⣿⣿⣶⣦⣤⣤⣀ ⠀⠀⠀⠀⢀⣤⣾⣿⣿⢫⡄⠀⠀⠀⠀⠀⠀⣿⣿⣹⣿⠏⢹⣿⣿⣿⣿⣿⣼⣿⠃⠀⠀⠀⢀⣿⡿⢀⣿⣿⠟⠀⠀⠀⠹⣿⣿⣿⠇⢿⣿⡄⠀⠀⠈⢿⣿⣷⣶⣶⣿⣿⣿⣿⣿⡿ ⣴⣶⣶⣿⣿⣿⣿⣋⣴⣿⣇⠀⠀⠀⠀⠀⢀⣿⣿⣿⣟⣴⠟⢿⣿⠟⣿⣿⣿⣿⣶⣶⣶⣶⣾⣿⣿⣿⠿⣫⣤⣶⡆⠀⠀⣻⣿⣿⣶⣸⣿⣷⡀⠀⠀⠸⣿⣿⣿⡟⠛⠛⠛⠉⠁⠀ ⠻⣿⣿⣿⣿⣿⣿⡿⢿⣿⠋⠀⢠⠀⠀⠀⢸⣿⣿⣿⣿⣁⣀⣀⣁⠀⠀⠉⠉⠉⠉⠉⠉⠉⠁⠀⠀⠀⠸⢟⣫⣥⣶⣿⣿⣿⠿⠟⠋⢻⣿⡟⣇⣠⡤⠀⣿⣿⣿⣿⡀⠀⠀⠀⠀⠀ ⠀⠀⠀ ⠉⠉⢹⣿⡇⣾⣿⠀⠀⢸⡆⠀⠀⢸⣿⣿⡟⠿⠿⠿⠿⣿⣿⣿⣿⣷⣦⡄⠀⠀⠀⠀⠀⠀⢠⣾⣿⣿⣿⣿⣯⣥⣤⣄⣀⡀⢸⣿⠇⢿⢸⡇⠀⢹⣿⣿⣿⡇⠀⠀⠀⠀⠀ ⠀⠀⠀ ⠀⠀⣾⣿⡇⣿⣿⠀⠀⠸⣧⠀⠀⢸⣿⣿⠀⢀⣀⣤⣤⣶⣾⣿⠿⠟⠛⠁⠀⠀⠀⠀⠀⠀⠀⠉⠉⠉⠙⠛⢛⣛⠛⠛⠛⠃⠸⣿⣆⢸⣿⣇⠀⢸⣿⣿⣿⣷⠀⠀⠀⠀⠀ ⠀⠀⠀⠀ ⠀⢻⣿⡇⢻⣿⡄⠀⠀⣿⡄⠀⢸⣿⡷⢾⣿⠿⠟⠛⠉⠉⠀⠀⠀⢠⣶⣾⣿⣿⣿⣿⣿⣶⣶⠀⠀⢀⡾⠋⠁⢠⡄⠀⣤⠀⢹⣿⣦⣿⡇⠀⢸⣿⣿⣿⣿⠀⠀⠀⠀⠀ ⠀⠀⠀ ⠀ ⠀⢸⣿⣇⢸⣿⡇⠀⠀⣿⣧⠀⠈⣿⣷⠀⠀⢀⣀⠀⢙⣧⠀⠀⠀⢸⣿⡇⠀⠀⠀⠀⢀⣿⡏⠀⠀⠸⣇⠀⠀⠘⠛⠘⠛⠀⢀⣿⣿⣿⡇⠀⣼⣿⢻⣿⡿⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀ ⠸⣿⣿⣸⣿⣿⠀⠀⣿⣿⣆⠀⢿⣿⡀⠀⠸⠟⠀⠛⣿⠃⠀⠀⢸⣿⡇⠀⠀⠀⠀⢸⣿⡇⠀⠀⠀⠙⠷⣦⣄⡀⠀⢀⣴⣿⡿⣱⣾⠁⠀⣿⣿⣾⣿⡇⠀⠀⠀⠀⠀ ⠀⠀ ⠀⠀⠀⠀⢻⣿⣿⣿⣿⣇⠀⢿⢹⣿⣆⢸⣿⣧⣀⠀⠀⠴⠞⠁⠀⠀⠀⠸⣿⡇⠀⠀⠀⠀⣿⣿⠀⠀⠀⠀⠀⠀⢀⣨⣽⣾⣿⣿⡏⢀⣿⣿⠀⣸⣿⣿⣿⡿⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠈⢻⣿⣿⣿⣿⣆⢸⡏⠻⣿⣦⣿⣿⣿⣿⣶⣦⣤⣀⣀⣀⣀⠀⣿⣷⠀⠀⠀⣸⣿⣏⣀⣤⣤⣶⣾⣿⣿⣿⠿⠛⢹⣿⣧⣼⣿⣿⣰⣿⣿⠛⠛⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠉⠛⠙⣿⣿⣦⣷⠀⢻⣿⣿⣿⣿⡝⠛⠻⠿⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡿⠿⠟⠛⠛⠉⠁⠀⠀⠀⣼⣿⣿⣿⣿⣿⣿⣿⠃⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠈⢻⣿⣿⣄⢸⣿⣿⣿⣿⣷⡄⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠙⠿⠟⠻⣿⡿⠋⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠙⢿⣿⣿⣿⣿⡌⠙⠛⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠈⠛⠛⠛⠃⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
An intuitive GUI application that automatically generates subtitle files (SRT) for videos using OpenAI's Whisper speech recognition model.
- User-friendly Interface: Clean design with intuitive controls
- Multiple Language Support: Auto-detection or specific language selection for 30+ languages
- Quality Options: Choose from different model sizes based on your needs
- Cross-Platform: Works on Windows, macOS, and Linux
- Progress Tracking: Real-time status updates and progress bar
- Detailed Logs: Track processing steps and troubleshoot issues
- Automatic FFmpeg Integration: No manual FFmpeg installation required
- Python 3.8 or higher
- Internet connection (for first-time model download)
git clone https://github.com/InboraStudio/Subtitle-Generator-AI.git
cd Subtitle-Generator-AIOr download and extract the ZIP file from GitHub.
python -m venv venv
venv\Scripts\activatepython3 -m venv venv
source venv/bin/activatepip install -r requirements.txtpython main.py-
Select a Video
- Click the "Browse" button next to "Video File"
- Choose your video file (.mp4, .mkv, .avi, etc.)
-
Choose Settings
- Model Size: Select based on your needs:
tiny: Fastest but less accurate (good for testing)base: Good balance between speed and accuracysmall: Better quality but slowermedium: High quality but slowerlarge: Best quality but slowest
- Language: Select a specific language or use "auto" for automatic detection
- Model Size: Select based on your needs:
-
Generate Subtitles
- Click the "Generate Subtitles" button
- The application will:
- Extract audio from your video
- Load the Whisper model (first run will download the model)
- Transcribe the audio
- Generate and save an SRT subtitle file
- Once completed, a success message will show the subtitle file location
-
Access Your Subtitles
- Click "Open Output Folder" to view the generated SRT file
- The subtitle file has the same name as your video file but with .srt extension
| Model | Memory Required | Speed | Accuracy | Recommended For |
|---|---|---|---|---|
| tiny | ~1GB | Fast | Basic | Quick testing, short clips |
| base | ~1GB | Good | Good | General purpose |
| small | ~2GB | Moderate | Better | More accurate transcription |
| medium | ~5GB | Slow | High | Detailed transcription |
| large | ~10GB | Very Slow | Highest | Professional transcription |
The application supports numerous languages, including but not limited to:
- English, Chinese, German, Spanish, Russian
- Korean, French, Japanese, Portuguese, Turkish
- Arabic, Italian, Hindi, Ukrainian, and many more
Auto-detection works well for most languages, but selecting a specific language can improve accuracy.
Solution: The application uses imageio-ffmpeg to automatically handle FFmpeg, but if you encounter errors:
- Check that imageio-ffmpeg is properly installed
- Restart the application
- If problems persist, install FFmpeg manually and add it to your PATH
Solution: This is normal if you don't have a compatible GPU. The application will use your CPU instead, which may be slower but will still work. For faster processing:
- Install PyTorch with CUDA support if you have a compatible NVIDIA GPU
- Use a smaller model size like "tiny" or "base"
Solution:
- Ensure you have an active internet connection
- If the download fails, restart the application
- Check your firewall settings if downloads are blocked
Solution:
- Use a smaller model size (tiny or base)
- Process shorter videos
- Close other resource-intensive applications
Solution:
- Ensure your video file isn't corrupted
- Try a different video format
- Check if the video file is accessible and not locked by another application
Solution:
pip install -r requirements.txtIf you encounter persistent issues:
- Check the detailed logs in the Log section of the application
- Restart the application and try with a different video file
- Try running with a smaller model size like "tiny" first
- Ensure you have sufficient disk space for model download and audio extraction
- Check system resources (CPU, RAM) while processing
subtitle-generator/
│
├── main.py # Main application entry point
├── subtitle_generator/ # Core application modules
│ ├── __init__.py # Package initialization
│ ├── gui.py # GUI components and event handlers
│ ├── processor.py # Video/audio processing logic
│ ├── whisper_patch.py # Whisper model integration with FFmpeg
│ └── utils.py # Utility functions
│
├── docs/ # Documentation assets
├── requirements.txt # Dependencies
├── LICENSE # MIT License
└── README.md # This documentation file
- whisper: OpenAI's speech recognition model
- tkinter: For GUI components
- imageio-ffmpeg: For automatic FFmpeg handling
- srt: For subtitle file generation
- numpy: For array operations
- torch: PyTorch for deep learning
The application uses a custom patch for Whisper's audio loading function to ensure seamless integration with FFmpeg, regardless of whether it's installed on the system or not. This is handled through:
- Automatic FFmpeg detection and environment setup
- Custom audio loading pipeline
- Proper error handling and reporting
To keep the UI responsive during processing, all heavyweight operations (audio extraction, model loading, transcription) run in separate threads. This allows for:
- Real-time status updates
- The ability to cancel ongoing operations
- A responsive interface even during lengthy transcriptions
This project is licensed under the MIT License - see the LICENSE file for details.
- OpenAI Whisper for the speech-to-text model
- FFmpeg for audio processing
- Inbora Studio for project development
Contributions are welcome! Please feel free to submit a Pull Request or open an Issue on the GitHub repository.