OCR in Docling

[aryan2000sp]

Hi,
I am new to using the Docling for document analysis. I want to know if I really need to ocr the whole document to get the layout analysis and table detection. Can I only use ocr when there is no text in pdf (scanned image)? I am asking this because I am using Docling to analyze the pdf document of 24 pages and it takes close to 70sec. I want to reduce this time to maybe 10-15sec.
I am also interested in hearing any different solutions?

[teowave]

Well, there are pdfs produced from scanned documents, and there are pdfs produced digitally, which usually include a text layer. For the digital ones, the ocr is not needed.