Implementation:Elevenlabs Elevenlabs python VoicePreviewResponseModel

Attribute	Value
Sources	src/elevenlabs/types/voice_preview_response_model.py
Domains	Voice Preview, Voice Design
Last Updated	2026-02-15

Overview

Description

The VoicePreviewResponseModel represents the response returned when generating a voice preview in the ElevenLabs SDK. It contains the base64-encoded audio data of the preview, along with the generated voice ID that can be used to create a permanent voice from the preview. The model also includes metadata such as the media type, duration, and language of the generated preview.

Usage

This model is returned by voice preview/design endpoints where a user generates a new voice preview before committing to creating a permanent voice. The generated_voice_id field can be used in subsequent API calls to finalize the voice creation.

Code Reference

Source Location

src/elevenlabs/types/voice_preview_response_model.py

Class Signature

class VoicePreviewResponseModel(UncheckedBaseModel):
    ...

Import Statement

from elevenlabs.types import VoicePreviewResponseModel

Base Class

UncheckedBaseModel (from elevenlabs.core.unchecked_base_model)

I/O Contract

Field	Type	Required	Description
audio_base_64	str	Yes	The base64 encoded audio of the preview.
generated_voice_id	str	Yes	The ID of the generated voice. Use it to create a voice from the preview.
media_type	str	Yes	The media type of the preview (e.g., audio/mpeg).
duration_secs	float	Yes	The duration of the preview in seconds.
language	Optional[str]	No	The language of the preview.

Usage Examples

import base64
from elevenlabs import ElevenLabs

client = ElevenLabs(api_key="your_api_key")

# Generate a voice preview (hypothetical endpoint usage)
preview = client.voice_design.generate_preview(
    text="Hello, this is a voice preview.",
    voice_description="A warm and friendly male voice",
    language="en"
)

print(f"Generated Voice ID: {preview.generated_voice_id}")
print(f"Media Type: {preview.media_type}")
print(f"Duration: {preview.duration_secs} seconds")
print(f"Language: {preview.language}")

# Decode and save the audio preview
audio_data = base64.b64decode(preview.audio_base_64)
with open("preview.mp3", "wb") as f:
    f.write(audio_data)

Related Pages

• Voice - The full voice model that can be created from a preview
• VoiceSamplePreviewResponseModel - Similar preview model for voice samples