Use this file to discover all available pages before exploring further.
A custom voice is a voice clone tied to your account. Once registered, it behaves exactly like a preset voice — same TTS endpoint, same parameters — but it lives at a separate set of endpoints and is only callable by the account that created it.Custom voices can be created two ways:
In Supertone Play — upload a sample and clone through the UI.
Via the API — POST /v1/custom-voices/cloned-voice with an audio file.
Either path produces the same kind of voice. Voices cloned in Play show up in list_custom_voices, and voices cloned via the API show up in Play. There’s no sync step.
Voice cloning via the API is not available on the Free tier. Cloning in Play is available on all paid tiers.
# List allresult = client.custom_voices.list_custom_voices(page_size=20)for voice in result.items or []: print(voice.voice_id, voice.name)# Search by name or descriptionresult = client.custom_voices.search_custom_voices(name="narrator")
// List allconst list = await client.customVoices.listCustomVoices({ pageSize: 20 });for (const voice of list.items ?? []) { console.log(voice.voiceId, voice.name);}// Search by name or descriptionconst search = await client.customVoices.searchCustomVoices({ name: "narrator" });
Once the clone is registered, call text_to_speech.create_speech exactly like you would with a preset voice — just pass the custom voice_id:
response = client.text_to_speech.create_speech( voice_id=CUSTOM_VOICE_ID, text="The first chapter begins on a quiet rainy morning.", language="en", model="sona_speech_2",)
Custom voices support the same voice_settings, output_format, include_phonemes, and normalized_text fields as preset voices.
Account-scoped. A custom voice is callable only by the account that created it. Calling someone else’s custom voice — even if you know the ID — returns 403 Forbidden.
Same credits. Custom voice calls deduct credits at the same rate as preset voices.
Permissions and disclosure. Make sure you have rights to clone any voice you upload. Check your jurisdiction’s rules around AI-generated voices and end-user disclosure.