Thanks! So if the PDF contains images with text within the image, it will convert the images to text with OCR (so like images of pages from a book). We do have an "Area Select" tool that lets you capture any screenshot within a page as a note, and in the future we're thinking of using GPT-4 vision where you can use that tool to select an image and ask questions about that image. This would be a cool feature indeed but not sure if users would actually use that very much in practice.