Keep the voice. Lift out the music.
An AI-powered application that isolates vocals from any audio or video file.
Voxa is a sophisticated audio processing tool wrapped in a sleek, premium, dark-mode user interface. Built with Python and Flask, it leverages Meta's state-of-the-art Demucs deep learning model to perform high-quality audio source separation.
Whether you upload an MP3 song or an MP4 video clip, Vocalizer intelligently strips away the instrumental backing tracks and returns a crystal-clear file containing only the isolated human vocals.
- Unrivaled AI Separation: Uses PyTorch and the Demucs model for industry-leading vocal extraction.
- Video & Audio Support: Seamlessly handles
.mp3,.wav,.flac,.mp4,.mkv, and more. If a video is uploaded, it extracts the audio, processes it, and stitches the isolated vocals back onto the original video file. - Premium UI/UX: A gorgeous, responsive, glass-inspired dark theme with drag-and-drop file support and dynamic loading states.
- Integrated Feedback: Built-in form submission for collecting user feedback directly to your inbox.
- Frontend: HTML5, Vanilla CSS (Custom Design System), JavaScript
- Backend: Python 3, Flask, Gunicorn
- AI & Processing: PyTorch, Meta's Demucs API, FFmpeg
Because this application runs heavy machine learning models, it is designed to be run locally or hosted on a dedicated server with sufficient RAM (4GB+ recommended).
- Python 3.9+ installed on your machine.
- FFmpeg installed and accessible in your system's PATH.
- Ubuntu/Debian:
sudo apt install ffmpeg - Mac:
brew install ffmpeg - Windows: Download via
winget install ffmpeg
- Ubuntu/Debian:
-
Clone the repository
git clone https://github.com/HassanKorey/Voxa.git cd Audio Filterer -
Create a virtual environment (Recommended)
python -m venv venv source venv/bin/activate # On Windows use `venv\Scripts\activate`
-
Install the dependencies
pip install -r requirements.txt
-
Run the application
python app.py
-
Open your browser and navigate to
http://127.0.0.1:7860. Drag and drop an audio file to see the magic happen!
- Demucs by Meta Research - The incredible AI model powering the audio separation.
- FormSubmit - Handling the frontend feedback form.