I use the following pipeline to generate the audio:
A C# unity file downloads a google spreadsheet as a .csv file. It then parses the spreadsheet and translates the higher level information in the spreadsheet into a text file with more practical details, such as what voice to use, what pace to say it at etc.
A shell script goes through the generated text file and uses the "say" command (available on Macs) to save the files into a folder.
No patreon. The purchasing of my game is all the patronage I want. :)