QuentinFuxa/WhisperLiveKit ? reverse-engineered prompt

Build me a self hosted app for live speech to text that I can run on my own machine or server. I want to open a local web page, talk into the mic, and see transcripts appear almost instantly as I speak, with support for multiple users at once. It should also be able to do speaker separation, translate speech into another language, and expose an API that feels compatible with OpenAI style transcription requests, plus a WebSocket option for real time streaming.

I’d like a simple command line way to start the server, choose a model, set the language, and transcribe audio files too, including subtitle output like SRT. If it makes sense, add basic model management, token protection for the API, and a clean demo page so it’s easy to test. Look up current docs online if you need to.