“In Tomek Porozynski's audiobook generation workflow, after the text is split by character, text-to-speech API requests can be sent in parallel to accelerate the process.”