VoiceStudio is an open-source desktop app for voice cloning, video dubbing, dictation, and transcription across 646 languages, running locally on your hardware with optional remote services. It supports NVIDIA and Apple Silicon acceleration, works on CPU-only machines, and lets you build custom voice models or use existing ones through a modular engine system.

• Runs entirely local by default with no cloud requirement; remote services and analytics are optional and require explicit consent
• Supports hardware acceleration on NVIDIA GPUs and Apple Silicon, but remains fully functional on older CPUs (though slower) with a ~5 GB PyTorch installation
• Switched from Tauri to Electron as the only desktop UI; integrates with coding agents via MCP and skills, and provides a local API for custom workflows
• Licensed under AGPL-3.0 for the app itself; voice cloning requires explicit permission and model licensing is handled separately from the application license