Joshua Oluseyi

Offline Neural Voice Lab

Offline Neural Voice Lab

Voice-to-voice conversion using a locally installed OpenVoice V2 inference worker. A reference voice can be assigned to a selected section of your own vocal or spoken recording. AI models run on your server, not through paid cloud API keys.

1. Register a permitted reference voice

2. Select the source recording and section

3. Converted audio

Worker status not checked.

Output is the selected passage only. Re-import the WAV into the Voice Swap & Vocal Sections editor or Recording Studio to construct a multivoice song. Conversion may change phrasing or tone and is not guaranteed to preserve singing articulation.

This requires locally installed neural weights, FFmpeg and an OpenVoice V2 worker. Shared hosting cannot usually install or run these models automatically. No voice should be impersonated without authorization.

Share this page by QROpen link