Generate spoken‑ready scripts from documents for podcasts, lectures, or summaries