VoxCPM2 · Local GPU

Clone Voice

Record or upload a clean voice sample, create the clone, then type any text to generate speech in the same voice.

Voice sample

10 seconds to 5 minutes · MP3, M4A or WAV · 20 MB max

Upload audioChoose a voice sample

Voice name

Preview text

Engine

VoxCPM2 · Local GPU

Text normalization

I confirm that I own this voice or have explicit permission to create and use a clone of it.

Type what the cloned voice should say

0/9999

Your generated voice appears here

The video generation workflows remain separate and unchanged.