Create Next App

Mode

Only needs a reference audio clip. No transcript required.

Voice (cloned)

Pick a voice from storage/reference_voices/

Drop <name>.wav (and optional <name>.txt for Studio mode) into storage/reference_voices/, then click Refresh.

Listing text

Each paragraph is rendered independently and stitched together — reliable for any length.

Natural-voice controls

Exaggeration (emotion intensity): 0.5Emotion intensity. 0 = flat / monotone. 0.5 = natural reading (default). 1.0 = very dramatic, theatrical.Pace control (lower = faster): 0.5How tightly the model follows reference pacing. Lower (0.3) = faster, more energetic. Higher (0.7) = slower, more deliberate. 0.5 is the sweet spot.Variation (lower = more consistent): 0.8Randomness in voice generation. Lower (0.5) = consistent but can sound flat. Higher (1.2) = more lifelike but less predictable. 0.8 is the documented sweet spot.

Exaggeration: 0 = flat, 1 = dramatic. Pace control: lower = faster delivery. Variation: lower = more deterministic. 0.5 / 0.5 / 0.8 is the documented sweet spot.

Music & SFX

Music bed

Upload

Background music

Whoosh SFX

Upload

Short SFX clip

Chime SFX

Upload

Short SFX clip

Mix levels (dB)

Voice gain: 0Music gain: -16Chime gain: -10Whoosh gain: -6

Timing & SFX placement

Paragraph pause (ms): 600Intro silence (ms): 800Outro silence (ms): 1200

Chime at each paragraph breakWhoosh at startWhoosh at end

Output

Output filename (optional, no extension)