Implementation:Elevenlabs Elevenlabs python Audio Output Utilities

Knowledge Sources	ElevenLabs Python
Domains	Audio_Processing, Media_Output
Last Updated	2026-02-15 00:00 GMT

Overview

Concrete tools for playing, saving, and streaming audio output provided by the elevenlabs-python SDK.

Description

The play.py module provides three standalone functions for consuming audio output from TTS generation:

• play(): Buffers audio and plays through ffplay (default), sounddevice, or IPython Audio
• save(): Writes audio bytes to a file
• stream(): Pipes audio chunks to mpv for progressive playback and returns the full audio

All three functions accept both bytes (complete audio) and Iterator[bytes] (streaming audio) as input.

Usage

Import these functions directly from the elevenlabs package. Use play for quick playback during development, save for production file generation, and stream when you need real-time playback of streaming TTS output.

Code Reference

Source Location

• Repository: elevenlabs-python
• File: src/elevenlabs/play.py
• Lines: L13-97

Signature

def play(
    audio: Union[bytes, Iterator[bytes]],
    notebook: bool = False,
    use_ffmpeg: bool = True,
) -> None:
    """Play audio through system speakers.

    Args:
        audio: Audio data or streaming audio chunks.
        notebook: Use IPython Audio display (for Jupyter).
        use_ffmpeg: Use ffplay for playback (default True).
            If False, uses sounddevice/soundfile.
    """

def save(
    audio: Union[bytes, Iterator[bytes]],
    filename: str,
) -> None:
    """Save audio to a file.

    Args:
        audio: Audio data or streaming audio chunks.
        filename: Output file path.
    """

def stream(
    audio_stream: Iterator[bytes],
) -> bytes:
    """Stream audio through mpv and return the full audio.

    Args:
        audio_stream: Streaming audio byte chunks.
    Returns:
        bytes: Complete concatenated audio data.
    """

Import

from elevenlabs import play, save, stream

I/O Contract

Inputs (play)

Name	Type	Required	Description
audio	Union[bytes, Iterator[bytes]]	Yes	Audio data or streaming chunks
notebook	bool	No	Use IPython Audio display (default False)
use_ffmpeg	bool	No	Use ffplay for playback (default True)

Inputs (save)

Name	Type	Required	Description
audio	Union[bytes, Iterator[bytes]]	Yes	Audio data or streaming chunks
filename	str	Yes	Output file path

Inputs (stream)

Name	Type	Required	Description
audio_stream	Iterator[bytes]	Yes	Streaming audio byte chunks

Outputs

Name	Type	Description
play() returns	None	Plays audio through speakers
save() returns	None	Writes audio file to disk
stream() returns	bytes	Full concatenated audio after progressive playback via mpv

Usage Examples

Play Audio

from elevenlabs import ElevenLabs, play

client = ElevenLabs()

audio = client.text_to_speech.convert(
    voice_id="JBFqnCBsd6RMkjVDRZzb",
    text="Hello world!",
    model_id="eleven_multilingual_v2",
)

# Play using ffplay (default)
play(audio)

# Play in Jupyter notebook
play(audio, notebook=True)

# Play using sounddevice (requires pip install sounddevice soundfile)
play(audio, use_ffmpeg=False)

Save to File

from elevenlabs import ElevenLabs, save

client = ElevenLabs()

audio = client.text_to_speech.convert(
    voice_id="JBFqnCBsd6RMkjVDRZzb",
    text="Save this to a file.",
    model_id="eleven_multilingual_v2",
    output_format="mp3_44100_128",
)

save(audio, "output.mp3")

Stream with mpv

from elevenlabs import ElevenLabs, stream

client = ElevenLabs()

audio_iterator = client.text_to_speech.convert(
    voice_id="JBFqnCBsd6RMkjVDRZzb",
    text="This will play progressively as it generates.",
    model_id="eleven_multilingual_v2",
)

# Plays in real-time via mpv and returns complete audio
full_audio = stream(audio_iterator)

Related Pages

Implements Principle

• Principle:Elevenlabs_Elevenlabs_python_Audio_Playback

Requires Environment

• Environment:Elevenlabs_Elevenlabs_python_FFmpeg_Mpv