There is no such method for this..

[polar heron]

Is there at least audio metadata with word timing? I saw there is something for websockets, but not sure I can find anything for generated files.

[fallow fiber]

yes it's possible but not on already generated audios..