The latest OCR API update doubles the performance of all our OCR API endpoints. In addition, it adds PDF autorotation. We already added image autoration last year. Today’s update expands this feature to PDF documents like invoices, receipts or FAX. You can test it on the online OCR form by checking “Detect Orientation”:
In the OCR API you can do the same with the new parameter detectOrientation. If set to true, the OCR engine looks at the PDF and, based on its text content, rotates the PDF page correctly. The rotation is done for each PDF page individually, which means it supports documents with pages of mixed orientation, too. In the API response the TextOrientation parameter is set for each page. It is either 0 (if the page is not rotated) or the degrees of the rotation. Autorotate increases the processing time slightly, so you should only use it if you need to.