Sopro
Local voice synthesis
Loading model
Preparing local runtime…
Drop a voice reference
Choose a file or record 5-20 seconds
Record
Choose
Text to speak
The smallest voice can still fill a room.
Stream
Offline
Backend
Language
Auto
EN
PT
FR
DE
Speak
↵
Voice controls
sampling
Temperature
0.80
Top P
0.90
Top K
Seed
Add a reference to begin.
0:00 / 0:00