A couple of improvement:
1. The app now can detect PDF language and switch to the corresponding default voice for Chinese and English (Ava for English and Lilian for Chinese).  You can still pick other voices on the list  as you wish.
2. Where to pause and where not to pause is much better optimized.
Something new:
If you have tried AI summary, this could be exciting. You can pick the language for AI summary, say summarizing in English for a Chinese PDF or vise versa. It means that you can upload a PDF in a language that you do not understand and let AI summarize for you in your own language. If your preferred language is not on the list or if you'd like to use a different API instead of Gemini, click App Support and send me an email. Notice that it still does not read PDFs with two-column text.