OCR
How to Make a Scanned PDF Searchable with OCR
7 min readUpdated October 7, 2026By the PDF Studio team
A scanned PDF looks like a document but behaves like a photograph: you cannot search it, select a sentence, or copy a figure out of it. Optical Character Recognition (OCR) fixes that by reading the page images and adding the text back as an invisible layer. Here is how to get a good result.
First: is your PDF actually scanned?
Many PDFs that look like scans already contain text. Open yours and try to select a word or press Ctrl/Cmd + F and search for one. If that works, the text layer exists and you do not need OCR. If nothing can be selected, every page is an image.
You can also run Extract Text: if it reports that no text was found, the PDF is image-only.
Start with a good scan
OCR accuracy depends mostly on the quality of the images you give it. A few habits make a large difference:
- Scan at 200 to 300 DPI. Lower is hard to read; higher just makes the file bigger.
- Keep pages straight and upright. Use Rotate PDF to fix pages that came out sideways or upside down before running OCR.
- Use even light and avoid shadows and creases if you are photographing pages.
- Black text on a plain light background works best. Faint ink, textured paper, and busy backgrounds reduce accuracy.
Step by step: run OCR
- 1Open OCR PDF and add your scanned PDF.
- 2Choose the language of the document. You can select up to three, which helps with documents that mix languages — for example English text with Spanish names. Choosing the wrong language produces gibberish.
- 3Run the recognition and download the result. The page looks exactly as before, but now has a searchable, selectable text layer behind the image.
- 4Check it: search for a word you know is on the page, and select a paragraph to see that the highlight lines up with the printed text.
When the results are poor
- Gibberish or missing letters: check the language setting, then rescan at 300 DPI and straighten the page.
- Handwriting: OCR is built for printed text. Handwriting is recognized unreliably, if at all.
- Tables: OCR reads the words but does not rebuild the table structure. Expect lines of text, and use the result as a starting point.
- Long documents: the 50-page limit means you should split bigger files with Split PDF and process the parts.
What to do with a searchable PDF
Once the text layer exists, a whole set of things becomes possible: copy passages, search across the document, extract all the text with Extract Text, turn it into an editable file with PDF to Word, or condense it with the AI Summarizer. If you plan to archive the document, run OCR first and then convert to PDF/A — see PDF/A explained.