🎙️ whisperx-batch-gui

Advanced Windows GUI for WhisperX with dynamic batch management, drag-and-drop queue, and speaker diarization support.

Topics: whisperx, gui, transcription, diarization, python-3-13, windows-app, batch-processing

A powerful graphical user interface for WhisperX, designed for batch transcription and speaker diarization of audio files. The project is optimized for Windows and supports dynamic queue management.

✨ Key Features

• 📦 Batch Processing: Add any number of audio files to the queue.
• ⚡ Dynamic Queue: Add or remove files directly while the processing is active.
• 🖱️ Drag-and-Drop: Easily reorder files in the queue by dragging (using the ☰ handle icon).
• 🗣️ Diarization: Automatic speaker identification (requires Hugging Face Token).
• 📝 Word-level Timestamps: Highly accurate timestamps for every single word.
• ⚙️ Flexible Settings:
  • Choose Whisper models (tiny, base, small, medium, large-v2, large-v3).
  • Configure diarization parameters (number of speakers, sensitivity threshold).
  • Manage batch size and chunk size to optimize for your VRAM.
• 📄 Multiple Output Formats: Save results in .srt, .txt, .json, and formatted text modes.

🚀 Installation

The project uses uv for automatic Python version management and dependency handling.

1. Download the repository.
2. Install FFmpeg:
   • Download from ffmpeg.org.
   • Add the bin folder path to your system's PATH environment variable.
3. Run the Installer:
   • Execute install.bat. It will automatically download the correct Python version (3.13), create a virtual environment, and install all dependencies.
4. Launch the App:
   • Use run.bat for daily usage.

🛠️ Hugging Face Setup

For diarization to work, you need to obtain an access token:

1. Create an account on Hugging Face.
2. Accept the model licenses (Accept License):
   • pyannote/segmentation-3.0
   • pyannote/speaker-diarization-3.1
3. Create a token in your profile settings (Access Tokens) and paste it into the application ("Diarization Settings" button).

🧪 Testing

The project includes an automated test suite to verify GUI logic and queue management.

To run tests:

python -m unittest test_dynamic_queue.py

Note: Tests include mocks for heavy libraries (torch, whisperx), so they can be run even without ML dependencies installed.

💻 Technical Details

• PyTorch 2.6+ Compatibility: The application includes specific fixes to allow loading model weights in recent PyTorch versions.
• Multi-threading: Audio processing runs in a background thread, keeping the GUI responsive.
• Caching: Hugging Face models are stored inside the .venv/cache folder, making it easy to keep your system clean.

---

Special Thanks To:

• oiik/win-gui-whisperx — for the original implementation ideas.
• Habr: Batch Transcription | Habr: WhisperX Guide