Why force users to create custom ai voice to swap out original audio voice? Just call it "Voice Swap" and once the trasncription and time stamping is done, pick an ai clone of a voice and regenerate audio.