You open a PDF, try to drag-select a paragraph, and nothing highlights. Ctrl+F finds no matches for a word you can plainly see on the screen. Copy and paste gives you nothing.
The document isn't broken and you're not doing it wrong. The PDF contains no text. It contains a picture of text — and to a computer, a picture of the word "invoice" is no more searchable than a photograph of a cat.
This guide covers turning that picture back into real, extractable text.
Why you can't select the text
There are two completely different kinds of PDF, and they look identical to you:
- Born-digital PDFs — exported from Word, a browser, or any program. These contain actual text objects. You can select, search, and copy them.
- Scanned PDFs — produced by a scanner, a phone camera, or a fax. Each page is an image. There are no letters in the file, just pixels that happen to look like letters.
Your eyes read both effortlessly. Software can only read the first.
The one-second test: try to select a word with your cursor. If nothing highlights, it's a scan, and you need OCR.
What OCR does
OCR — Optical Character Recognition — looks at the image, recognises the shapes as letters, and writes a text layer into the PDF. The layer is invisible; it sits precisely behind the picture of each word.
The result looks exactly the same on screen, but now:
- Ctrl+F finds your words.
- You can select and copy text.
- The document is indexable by search engines and document-management systems.
- It can be converted to Word or Excel with meaningful results.
The page still looks like a scan. It just isn't dumb any more.
The steps
Open Blackpdf's OCR PDF tool and drop your scanned file in.
Select the language of the document. This matters more than people expect — see our guide on OCR in other languages.
Click Start OCR and wait — recognition takes longer than most PDF operations, because the engine is genuinely reading every page.
Download the result and test it: open it and Ctrl+F for a word you can see. If it's found, it worked.
Get better results: clean the image first
OCR accuracy is overwhelmingly a function of image quality, not of the engine. Recognition works by matching shapes to letterforms — so anything that blurs the line between ink and paper hurts it: shadows, grey backgrounds, soft focus, skew.
If your scan is a phone photo or an old, murky document, run it through Enhance PDF first. Flattening the lighting and raising the contrast can turn garbage recognition into near-perfect recognition, with no change to the OCR settings at all.
The order that matters:
- Rotate — get the page upright (sideways pages are catastrophic for OCR).
- Enhance — clean the image.
- OCR — now recognise the text.
- Compress — last. Never before OCR, because compression degrades exactly the letterforms the engine reads.
The full reasoning is in cleaning up a scan before OCR.
Common questions
Will OCR make the page look different?
No. The image is unchanged — OCR adds an invisible text layer behind it. The document looks exactly as it did; it's just searchable now.
Is the extracted text always accurate?
No. OCR is very good, not perfect. Accuracy depends on image quality, font,
and language. Clean, high-contrast printed text gets near-perfect results;
handwriting, ornate fonts, and poor scans get errors. Always spot-check
anything that matters — especially numbers, where a 0/O or 1/l
confusion changes the meaning.
Can it read handwriting?
Generally, no. OCR is built for printed type. Neat block capitals sometimes work; cursive essentially never does. Handwriting recognition is a different and much harder problem.
How do I get the text out into a document I can edit?
Once the PDF is OCR'd, convert it — PDF to Word for a document, or PDF to Excel for tables. Converting a scan without OCR gives you a Word file containing a picture, which is useless.
Does OCR make my file bigger?
Slightly — you're adding a text layer. It's a small addition, and it does not shrink the file (a common misconception). If size matters, compress it afterwards.
My PDF is a mix of real text and scanned pages.
Running OCR is still fine — the pages that already have text keep it, and the image-only pages gain a text layer.
Wrap-up
- Test it: can't select the text? It's a scan.
- Clean it first with Enhance PDF if it's a phone photo or a murky old document.
- Run OCR PDF, choosing the right language.
- Verify with Ctrl+F.
A scan is a picture of words. OCR turns it back into words — and once it is text, everything else (searching, copying, converting to Word) becomes possible.