Feature Request: Realtime Websocket output with sentence end event

[EuphieEuphoria]

Hello! I stumbled across this project from a friend while integrating something similar into Resonite. Currently I am using an implementation of speech to text that uses google's speech API which is prone to text errors. Your implementation solves many of the issues I have had with google's API.

Currently, Resonite supports websocket events by receiving and sending websocket messages. My current implementation live streams the generated text over the websocket connection and finalizes the sentence by sending a message containing [speechEnded]. I am using my own visual code to parse the websocket and display the text in game.

If TaSTT had a similar implementation of websockets, I could completely forgo using google's speech to text API altogether.

My request is the implementation of a locally hosted websocket that broadcasts the text messages generated live and completes the sentence with a dedicated end text message.

(Example in-game receiver I am using currently)