Just like the last update (#62) it would be great to also send utterance and speak messages to the gui web socket so gui clients can use it to display the transcribed + spoken text. For confirmation of the utterances for instance. Also better for accessibility for deaf and hardhearing people or for users with temporary sound off.