# Websockets for Responses API

**URL:** <https://community.openai.com/t/websockets-for-responses-api/1374906>\
**Category:** API\
**Tags:** api, responses-api\
**Created:** [February 23, 2026, 8:48pm UTC](https://community.openai.com/t/websockets-for-responses-api/1374906 "2026-02-23T20:48:13Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![platypus](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/platypus/32/759972_2.png) [@platypus](https://community.openai.com/u/platypus)\
**Post date:** [February 23, 2026, 8:48pm UTC](https://community.openai.com/t/websockets-for-responses-api/1374906/1 "2026-02-23T20:48:13Z")

</div>

Looks like `openai` `v2.22.0` lib now has websockets implementation for the Responses API, which should significantly improve the time-to-first-token, [Release v2.22.0 · openai/openai-python · GitHub](https://github.com/openai/openai-python/releases/tag/v2.22.0). On Codex-Spark it was 50% improvement so let’s see.

---

<div class="post-metadata">

**Author:** ![aprendendo.next](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/aprendendo.next/32/721758_2.png) [@aprendendo.next](https://community.openai.com/u/aprendendo.next)\
**Post date:** [February 23, 2026, 10:27pm UTC](https://community.openai.com/t/websockets-for-responses-api/1374906/2 "2026-02-23T22:27:41Z")

</div>

Seems to be good for tool calling.

> **[WebSocket Mode | OpenAI API](https://developers.openai.com/api/docs/guides/websocket-mode/)**
>
> Learn how to use Responses API WebSocket mode with response.create and previous\_response\_id.