Voice

Voice Cloning

Instant voice cloning from samples, ElevenLabs shape. Cloned voices work in every text-to-speech route.

Last reviewed 2026-08-25

Clone a voice

POST /elevenlabs/v1/voices/add with one or more samples. The first sample is the reference. The response is the ElevenLabs voice object with your new voice_id.

from elevenlabs.client import ElevenLabs

client = ElevenLabs(base_url="https://api.mserve.ai/elevenlabs", api_key="ms_live_your_key")

voice = client.voices.ivc.create(
    name="Studio voice",
    files=[open("sample.wav", "rb")],
    description="Warm narration, consented 2026-08-25",
    labels='{"use": "narration"}',
)
print(voice.voice_id)

Prop

Type

Use the voice

Pass the returned voice_id to any text-to-speech route. Cloned voices use chatterbox and its per-character price.

audio = client.text_to_speech.convert(voice_id=voice.voice_id, text="Every model. One API.")

Manage voices

RoutePurpose
POST /elevenlabs/v1/voices/{voice_id}/editRename, add samples, change description or labels.
POST /elevenlabs/v1/voices/{voice_id}/settings/editUpdate voice settings.
DELETE /elevenlabs/v1/voices/{voice_id}Remove the voice and its samples.
GET /elevenlabs/v1/voices/{voice_id}/samples/{sample_id}/audioDownload a sample.
DELETE /elevenlabs/v1/voices/{voice_id}/samples/{sample_id}Remove one sample.

Consent

Clone only voices you have permission to use. Samples stay with your account and are deleted with the voice.

On this page

Share feedback