Skip to main content

​
Overview

The voices.ivc.create() method creates an instant voice clone (IVC) and adds it to your collection of voices. IVC allows you to clone a voice from audio samples in seconds.
IVC (Instant Voice Cloning) creates a voice clone quickly from audio samples. For higher quality clones with more customization options, consider using PVC (Professional Voice Cloning).

​
Method Signature

client.voices.ivc.create(
    name: str,
    files: List[core.File],
    remove_background_noise: Optional[bool] = None,
    description: Optional[str] = None,
    labels: Optional[IvcCreateRequestLabels] = None,
    request_options: Optional[RequestOptions] = None
) -> AddVoiceIvcResponseModel

​
Parameters

​
Returns

​
AddVoiceIvcResponseModel
object
Response containing the newly created voice information.

​
Examples

​
Basic Voice Clone

Create a voice clone from a single audio file: 
from elevenlabs import ElevenLabs

client = ElevenLabs(api_key="YOUR_API_KEY")

with open("sample.mp3", "rb") as f:
    response = client.voices.ivc.create(
        name="My Cloned Voice",
        files=[f]
    )

print(f"Voice cloned successfully!")
print(f"Voice ID: {response.voice_id}")
print(f"Requires verification: {response.requires_verification}")

​
Clone with Multiple Samples

Create a voice clone from multiple audio samples for better quality: 
from elevenlabs import ElevenLabs

client = ElevenLabs(api_key="YOUR_API_KEY")

audio_files = ["sample1.mp3", "sample2.mp3", "sample3.mp3"]

with open(audio_files[0], "rb") as f1, \
     open(audio_files[1], "rb") as f2, \
     open(audio_files[2], "rb") as f3:
    
    response = client.voices.ivc.create(
        name="Professional Voice Clone",
        files=[f1, f2, f3],
        description="Voice clone for professional narration"
    )

print(f"Voice ID: {response.voice_id}")

​
Clone with Background Noise Removal

Create a voice clone and remove background noise: 
from elevenlabs import ElevenLabs

client = ElevenLabs(api_key="YOUR_API_KEY")

with open("noisy_sample.mp3", "rb") as f:
    response = client.voices.ivc.create(
        name="Clean Voice Clone",
        files=[f],
        remove_background_noise=True,
        description="Cloned from noisy audio sample"
    )

print(f"Voice ID: {response.voice_id}")

​
Clone with Labels

Create a voice clone with descriptive labels: 
from elevenlabs import ElevenLabs

client = ElevenLabs(api_key="YOUR_API_KEY")

with open("british_voice.mp3", "rb") as f:
    response = client.voices.ivc.create(
        name="James - British Narrator",
        files=[f],
        description="British male voice for audiobook narration",
        labels={
            "language": "en",
            "accent": "british",
            "gender": "male",
            "age": "middle-aged"
        }
    )

print(f"Voice created: {response.voice_id}")

​
Use Cloned Voice for Text-to-Speech

Create a voice clone and immediately use it for generation: 
from elevenlabs import ElevenLabs

client = ElevenLabs(api_key="YOUR_API_KEY")

# Clone the voice
with open("sample.mp3", "rb") as f:
    clone_response = client.voices.ivc.create(
        name="My Voice",
        files=[f]
    )

voice_id = clone_response.voice_id
print(f"Voice cloned: {voice_id}")

# Use the cloned voice for text-to-speech
if not clone_response.requires_verification:
    audio = client.text_to_speech.convert(
        voice_id=voice_id,
        text="Hello! This is my cloned voice speaking."
    )
    
    # Save the audio
    with open("output.mp3", "wb") as f:
        f.write(audio)
    
    print("Audio generated with cloned voice!")
else:
    print("Voice requires verification before use.")

​
Async Voice Cloning

Create a voice clone asynchronously: 
import asyncio
from elevenlabs import AsyncElevenLabs

client = AsyncElevenLabs(api_key="YOUR_API_KEY")

async def main():
    with open("sample.mp3", "rb") as f:
        response = await client.voices.ivc.create(
            name="Async Cloned Voice",
            files=[f],
            description="Voice cloned asynchronously"
        )
    
    print(f"Voice ID: {response.voice_id}")
    print(f"Requires verification: {response.requires_verification}")

asyncio.run(main())

​
Best Practices

  • Use lossless or high-bitrate audio formats (WAV, FLAC, or 320kbps MP3)
  • Ensure samples are free from compression artifacts
  • Remove silence from the beginning and end of samples
  • Aim for consistent volume levels across samples
  • Include diverse emotional tones and speech patterns
  • Use samples with clear pronunciation
  • Avoid samples with overlapping voices or music
  • Provide at least 1-2 minutes of total audio
  • Some voices may require verification to prevent misuse
  • Ensure you have proper rights to clone the voice
  • Follow ElevenLabs’ terms of service for voice cloning

Build docs developers (and LLMs) love

Get started for freeTalk to us