🎙️ LuxTTS Voice Cloning

Note: This demo runs on a 2-core CPU, so expect slower inference.
Tip: If words get cut off, lower Speed or increase Ref Duration.

Text to Synthesize

Reference Audio (.wav)

RMS (Loudness)

Reference Duration (sec)

Lower = faster. Set ~1000 if you hear artifacts.

T-Shift

Num Steps

1 10

Speed (Lower = Longer / Clearer)

0.5 2

Return Smooth

Result

Ready to generate...