Voice Cloning

Create custom voice clones from audio samples, then use them with TTS to generate speech in that voice.

Clone limits

FreeProPremium1 clone5 clones20 clones

Typical workflow

Voice cloning is a two-step process: prepare a sample, then create the clone.

End-to-end example

const { sample, error: sampleErr } = await sonus.voices.prepareSample({
  url: 'https://example.com/voice-sample.mp3',
})
if (sampleErr) throw sampleErr

const { voice, error: voiceErr } = await sonus.voices.create({
  name: 'My Voice',
  audio: sample.audio,
  format: sample.format,
  sampleRate: sample.sampleRate,
})
if (voiceErr) throw voiceErr

const { sound, error: ttsErr } = await sonus.tts.generate({
  text: 'Hello in my cloned voice!',
  voiceId: voice.id,
  provider: 'fish',
})

sonus.voices.prepareSample({ url })

Download and prepare an audio sample for voice cloning. Audio must be 3\u201340 seconds.

Parameter	Type	Description
`url`required	`string`	URL to an audio file (3–40 seconds)

const { sample, error } = await sonus.voices.prepareSample({
  url: 'https://example.com/voice-sample.mp3',
})

sonus.voices.create({ name, audio, format, sampleRate, skipAsr? })

Create a voice clone from a prepared audio sample.

Parameter	Type	Description
`name`required	`string`	Name for the voice clone
`audio`required	`string`	Base64-encoded audio from prepareSample
`format`required	`string`	Audio format from prepareSample
`sampleRate`required	`number`	Sample rate from prepareSample
`skipAsr`	`boolean`	Skip automatic speech recognition (default: false)

const { voice, error } = await sonus.voices.create({
  name: 'My Custom Voice',
  audio: sample.audio,
  format: sample.format,
  sampleRate: sample.sampleRate,
})

sonus.voices.list()

List all your voice clones.

const { voices, error } = await sonus.voices.list()

sonus.voices.delete(id)

Delete a voice clone by ID.

const { ok, error } = await sonus.voices.delete('voice_abc123')