New open source model makes voice cloning easy, and free—on practically any computer: https://www.youtube.com/watch?v=dQ841Pd6YvQ
Could this be used to read pdfs?