How to make a scanned PDF searchable
7 min read
A scanned page is a picture of text, so you cannot search it, select a sentence or copy a figure out of it. OCR, optical character recognition, reads the picture and adds the text it finds as an invisible layer on top of the page. The page looks the same and behaves like a normal document.
What you end up with
- The same page image, with an invisible text layer positioned over each word.
- Ctrl+F finds words, and you can select and copy text in any PDF viewer.
- The layer is only as good as the recognition, so some words will be wrong. Check numbers and names against the page.
Steps
- Open OCR a scanned PDF and choose the language the document is written in.
- Leave cleaning on unless the scan is already crisp and white.
- Choose your PDF. Recognition runs in your browser, one page after another, with a progress estimate.
- Search the results screen for a word you know is in the document, then download the new PDF.
- Open the PDF and press Ctrl+F to confirm it is searchable in your own viewer.
What decides accuracy
- Resolution. About 300 DPI is the usual target for text. Small type scanned at a low resolution has too few pixels per letter.
- Contrast and cleanliness. Recognition engines are trained on clean, dark text on a plain background. Faint ink, a grey cast and shadows cost accuracy.
- Straightness. A page scanned at an angle reads less accurately.
- Language. The engine needs the right language model. Here that is English, French, German or Spanish.
- Content. Printed text reads well. Handwriting, unusual fonts and dense small print do not.
Why clean the scan first
On this site’s faded demo scan, the recognition engine found 6 of 24 test words in the original and 18 of 24 after cleaning, and it ran faster on the cleaned page. That is one synthetic English sample, so read it as an illustration, not a guarantee. The reason is general: engines are trained on high-contrast text, and a clean page is closer to what they expect.
That is why the sensible order is scan well, enhance, then recognise. The OCR tool does the cleaning step for you.
Privacy: where does the recognition run?
Many OCR services upload your document to a server. For contracts, medical records or anything confidential, that matters. The OCR tool here runs the recognition engine in your browser, from files served by this site, so the document is not sent to another server. The tradeoff is speed: it runs on your own device, so expect several seconds per page, more on a phone.
What OCR will not do
- It does not read handwriting.
- It does not rebuild tables or layout; the text layer is words in position, which is what search and copy need.
- It does not fix a bad scan. If the page is unreadable to you, it will be unreadable to the engine.
Questions
Does my PDF already have text?
Select a sentence with the mouse. If a neat highlight follows the words, the page has real text. If a rectangle sweeps across the page, it is an image, and OCR is what you need.
Why are some words wrong?
Recognition is a statistical guess. Faint, small, skewed or unusual text is the usual cause. Check figures and names against the page, and rescan at 300 DPI if the original is available.
Keep reading
- How to Improve Scanned PDF QualityMost scan quality is decided at capture time. Resolution, colour mode, lighting and compression settings that make every later step easier.
- How to Clean Up a Scanned PDFRemove grey backgrounds, speckle, shadows and show-through from scanned PDFs — and keep the faint content you actually wanted.
- How to Make a Scanned Document More ReadableReadability is more than contrast: resolution, sharpening, colour handling, OCR and accessibility all decide whether a scan is genuinely usable.
Try it on your own file
Everything described above runs in your browser — nothing is uploaded.