Have integrated openedai-speech, and now realized that an amazing change would be to have the ability to set specific voices when creating new models.