Implementation:Openai Openai python Websocket Connection Options

Knowledge Sources	Openai_Openai_python OpenAI API Reference
Domains	API_Types, Python
Last Updated	2026-02-15 00:00 GMT

Overview

Concrete tool for configuring WebSocket connection options when using the OpenAI Realtime API, provided by the openai-python SDK.

Description

WebsocketConnectionOptions is a TypedDict that exposes WebSocket connection configuration options sourced from the websockets library. It allows fine-tuning of connection parameters including extensions (negotiated WebSocket extensions), subprotocols (preferred subprotocols), compression (permessage-deflate control), max_size (incoming message size limit), max_queue (receive buffer high-water mark), and write_limit (write buffer high-water mark). These options are passed through to the underlying websockets.asyncio.client.connect call.

Usage

Import this type when you need to customize the WebSocket connection parameters for the OpenAI Realtime API, particularly for performance tuning, protocol negotiation, or disabling compression.

Code Reference

Source Location

• Repository: openai-python
• File: src/openai/types/websocket_connection_options.py

Signature

class WebsocketConnectionOptions(TypedDict, total=False):
    """Websocket connection options copied from websockets."""

    extensions: Sequence[ClientExtensionFactory] | None
    subprotocols: Sequence[Subprotocol] | None
    compression: str | None
    max_size: int | None
    max_queue: int | None | tuple[int | None, int | None]
    write_limit: int | tuple[int, int | None]

Import

from openai.types import WebsocketConnectionOptions

I/O Contract

Fields

Name	Type	Required	Description
extensions	Sequence[ClientExtensionFactory] or None	No	List of supported extensions, in order of negotiation and execution.
subprotocols	Sequence[Subprotocol] or None	No	List of supported subprotocols, in order of decreasing preference.
compression	str or None	No	Compression setting. Set to None to disable permessage-deflate (enabled by default).
max_size	int or None	No	Maximum size of incoming messages in bytes. None disables the limit.
max_queue	int or None or tuple[int or None, int or None]	No	High-water mark of the buffer where frames are received. Defaults to 16 frames.
write_limit	int or tuple[int, int or None]	No	High-water mark of write buffer in bytes. Defaults to 32 KiB.

Usage Examples

from openai.types import WebsocketConnectionOptions

ws_options: WebsocketConnectionOptions = {
    "compression": None,       # Disable compression
    "max_size": 10 * 1024 * 1024,  # 10 MB max message size
    "max_queue": 32,           # 32 frames buffer
}

# Pass to Realtime API connection
# client.beta.realtime.connect(model="gpt-4o-realtime", **ws_options)

Related Pages

• Environment:Openai_Openai_python_Python_3_9_Plus