About Splitvoice
Empowering creators with AI-powered audio tools that respect privacy.
Our Story
Splitvoice was born from a simple frustration: existing stem separation tools either required uploading your music to the cloud (sacrificing privacy), were prohibitively expensive, or produced mediocre results.
We built Splitvoice to bridge the gap between professional-grade separation and local-first privacy. By leveraging state-of-the-art neural networks from Meta AI and the open-source community, we deliver studio-quality stem separation that runs entirely on your machine.
No subscriptions. No data uploads. No compromises.
Our Values
We believe in:
- Privacy First: Your audio stays on your computer. Period.
- Open Technology: Built on open-source AI models like Demucs and BS-RoFormer
- One-Time Purchase: Fair pricing without recurring fees
- Quality Above All: Fine-tuned models for the best possible separation
Built With
Splitvoice is powered by:
- Meta Demucs - Hybrid Transformer model for music source separation
- BS-RoFormer - Band-split rotary transformer for instrument separation
- Electron - Cross-platform desktop framework
- PyTorch - Deep learning framework for AI inference
- Python - Core processing backend