How to Extract Text From a Scanned PDF
A scanned PDF looks like text but contains none - it's an image. Here's the real workflow to actually get text back out of it.
A scanned PDF looks exactly like a normal document, but there's no actual text underneath it - just a picture of a page, stored inside a PDF wrapper. Selecting or copying text from one does nothing, because as far as the file format is concerned, there's no text there to select.
The workflow
- Convert the scanned PDF's pages to images - JPG or PNG both work.
- Run OCR on each page image to recognize the text.
- Copy the recognized text out for each page.
- Combine the text from all pages into one document if the source PDF had more than one page.
- Proofread the result against the original scan, especially anything with numbers, names, or unusual formatting - these are the most error-prone for OCR.
Why this is page by page rather than one step
This workflow chains two separate tools - image conversion and OCR - rather than using one dedicated PDF-OCR tool that does both in a single pass. It takes a few more steps for a multi-page document, but it's a reliable path when no all-in-one PDF-OCR tool is available.
When the result doesn't need to be perfect
If you just need the general content of a scanned document - to search for a keyword, or get the gist without retyping the whole thing - OCR output that's 90-95% accurate is usually good enough. Save the careful proofreading step for documents where every word actually matters, like a contract or a form being retyped for official use.
Frequently asked questions
Why can't I just select and copy text from a scanned PDF like I can with a normal one?
A scanned page is stored as an image, not as actual text characters - there's nothing for your PDF viewer to select, no matter how much it visually looks like readable text.
Can I extract text from every page of a scanned document at once?
Without a dedicated batch PDF-OCR tool, it's a page-by-page process: convert each page to an image, then run OCR on each one individually. It's more steps than a single-click tool, but it works reliably with standard image and OCR tools.
Will the extracted text need manual corrections?
Often, at least a little - OCR accuracy depends heavily on scan quality, so proofreading the result against the original, especially for anything important, is worth the few extra minutes.