I'm trying the new RecognizeDocumentsRequest supposed to detect paragraphs (among other things) in a document.
I tried many source images, and I don't see the slightest difference compared to the old API (VN)RecognizedTextRequest
Is it supposed to not work or is it in beta?