I am also curious about this. ADI is reliable but does have edge case issues on malformed PDF

I fear tesseract OCR is a potential limitation though. I’ve seen it make so many mistakes