Implementation:Openai Openai python Local Audio Player

Knowledge Sources	openai-python
Domains	Audio, Client_Helpers
Last Updated	2026-02-15 00:00 GMT

Overview

Concrete helper class for playing audio through the local sound device provided by the OpenAI Python SDK.

Description

The LocalAudioPlayer class provides play() for complete audio playback and play_stream() for streaming audio playback. It accepts TTS responses (HttpxBinaryResponseContent), numpy arrays, and async audio generators. It uses the sounddevice library for audio output.

Usage

Instantiate with an optional should_stop callback. Call play() with a TTS response or numpy array, or play_stream() with an async generator of audio chunks.

Code Reference

Source Location

Repository: openai-python
File: src/openai/helpers/local_audio_player.py
Lines: L19-165

Signature

class LocalAudioPlayer:
    def __init__(
        self,
        should_stop: Callable[[], bool] | None = None,
    ) -> None:
        """
        Args:
            should_stop: Optional callback returning True to stop playback.
        """

    async def play(
        self,
        input: Union[
            npt.NDArray[np.int16],
            npt.NDArray[np.float32],
            HttpxBinaryResponseContent,
            AsyncStreamedBinaryAPIResponse,
            StreamedBinaryAPIResponse,
        ],
    ) -> None:
        """Plays audio from various input types."""

    async def play_stream(
        self,
        buffer_stream: AsyncGenerator[Union[npt.NDArray[Any], None], None],
    ) -> None:
        """Plays audio from a streaming generator."""

Import

from openai.helpers import LocalAudioPlayer

I/O Contract

Inputs

Name	Type	Required	Description
input	NDArray or HttpxBinaryResponseContent	Yes (play)	Audio data to play
buffer_stream	AsyncGenerator	Yes (play_stream)	Async generator of audio chunks
should_stop	Callable[[], bool]	No	Stop predicate for interruption

Outputs

Name	Type	Description
(audio playback)	None	Audio played through local sound device

Usage Examples

Play TTS Response

import asyncio
from openai import OpenAI
from openai.helpers import LocalAudioPlayer

client = OpenAI()
response = client.audio.speech.create(
    model="tts-1",
    voice="alloy",
    input="Hello from the OpenAI SDK!",
    response_format="pcm",
)

player = LocalAudioPlayer()
asyncio.run(player.play(response))

Streaming Playback with Stop Control

import asyncio

stop_flag = False

def should_stop():
    return stop_flag

player = LocalAudioPlayer(should_stop=should_stop)

with client.audio.speech.with_streaming_response.create(
    model="tts-1",
    voice="nova",
    input="A long text to be read aloud.",
    response_format="pcm",
) as response:
    asyncio.run(player.play(response))

Related Pages

Implements Principle

Principle:Openai_Openai_python_Audio_Playback

Requires Environment

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment