Use [brackets] to add audio tags for supported models. e.g. [laughs], [sighs], [whispers]