Feature Request: Implement OpenAPI's realtime streaming speech to text protocol

[candrews]

Prerequisites

• I am running the latest code. Mention the version if possible as well.
• I carefully followed the README.md.
• I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
• I reviewed the Discussions, and have a new and useful enhancement to share.

Feature Description

OpenAI specifies a protocol for realtime speech to text streaming transcription streaming with wss://<server>/v1/realtime?intent=transcription, see "Streaming the transcription of an ongoing audio recording" at https://developers.openai.com/api/docs/guides/speech-to-text#streaming-the-transcription-of-an-ongoing-audio-recording

The implementation of OAI STT in llama.cpp, added in #21863 by @ngxson, is for the http /v1/audio/transcriptions protocol. This feature request is for adding realtime streaming support using the websocket protocol.

Motivation

By implementing the OpenAI streaming STT protocol, many existing clients will be able to leverage this feature to provide more responsive realtime transcription of live audio. For example, I'm particularly interested in enabling Wyoming OpenAI's realtime streaming support which will improve the responsiveness of Home Assistant voice assistants as documented at roryeckel/wyoming_openai#59

Possible Implementation

No response

[06kellyjac]

@candrews does the title need amending? it mentions "text to speech" but the content refers to speech to text