Features Scene-based generation with request stitching for consistent prosody across scenes Character presets (narrator, salesperson, expert, dramatic, calm) for natural delivery Single scene ...