Voice Clone

70 credits

Reference voice

No audio selected

Source

Reference voice audio

Use a clear 5-30 second clip. Longer files use the first 30 seconds.

Cloned output

Generated audio appears here

Run the tool to create cloned speech.

Reference voice

Drop or click to upload audio.

Clear 5-30s clips work best; longer files use the first 30s.

Speech text67/1200

Emotion tags such as [excited], [whisper], or [speaking slowly] intentionally change the delivery. Remove them for neutral speech.

Voice settings

LanguageSet this manually if auto-detection mispronounces the script.Consistency · temperature0.7Lower values are steadier and usually sound more natural; higher values add emotion and variation.Variation · top P0.7Lower this when the voice wanders, slurs, or changes delivery unexpectedly.Repetition penalty1.2Raise slightly for stutters or repeated words; too high can hurt pronunciation.Chunk length200200 is balanced. Lower can help short phrases; higher can improve continuity in longer speech.Maximum acoustic tokensLeave at 0 for automatic length. Increase only if long speech is cut off.