Planning to finetune open source tts model on my voice and see if I get it to naturally implement things like laughs through tags. The voice im using in vapi implements breaths so I know it possible ...just probably difficult