Since you are already integrating all these different AI tools, why not add one more (like ElevenLabs) for additional voice options? Text-to-speech often loses the inflection of what was intended, but I find using ElevenLabs Voice Changer keeps the inflection intact while offering a multitude of different voices and languages.
Thank you for the suggestion. We are evaluating this for a future update. 
Great Idea. If the voice is off only a little, the whole effort is in the sewer!