Low latency ai companion voice talk in 60 lines of code using faster_whisper and elevenlabs input streaming - View it on GitHub