How to make a scanned PDF searchable

To make a scanned PDF searchable, open it at pdfviewer.work, click OCR and wait while the text is recognized in your browser; then press Ctrl+F to confirm and click Download. The recognized text is stored as an invisible layer under the scan, so the page looks the same but search and copy now work.

Make a scanned PDF searchable in five steps

1. Open the scan. Drop the PDF here, on the OCR tool page, or click Open a PDF.
2. Click OCR. Recognition runs page by page in the browser tab using tesseract.js. The language follows your browser's language setting, with English as the default.
3. Wait. Expect tens of seconds per page on a laptop; a 30-page scan is a coffee break, not a click. Leave the tab open.
4. Test it. Press Ctrl+F and type a word you can see on the page. If the match is found, the text layer is in place. Try selecting and copying a line as a second check.
5. Click Download. The new PDF contains the invisible text layer; other readers will search it too.

Why a scanned PDF is not searchable to begin with

A scanner produces a photograph of each page. The PDF wrapper around those photographs has no idea what the pixels say, which is why Ctrl+F finds nothing and copy-paste returns nothing. A PDF from Word, by contrast, stores every character as text. The scanned PDF page goes deeper into the two kinds of PDF.

OCR (optical character recognition) reads the picture and writes a text layer that is positioned under each word but drawn invisibly. The scan stays exactly as it looked; what changes is that search, copy, screen readers and indexing now have something to work with. The OCR explainer covers how the engine decides what a letter is.

Other ways to OCR a PDF on Mac, Windows and the command line

If you only need the words and not the PDF, OCR in the viewer and then copy the text out; the PDF to text page explains the ways to extract plain text.

Before and after OCR: preparing the scan and using the result

A little preparation improves recognition. Rotate sideways pages from the thumbnail sidebar so the text reads the right way, delete blank backs and cover sheets so the engine does not spend time on them, and if several scans belong together, merge them first so you run OCR once. All of these are per-page controls in the sidebar; the search inside a PDF page shows how search behaves once a text layer exists.

After OCR, the scan behaves like a born-digital PDF for most purposes: Ctrl+F with next and previous match, selecting and copying a paragraph into an email, highlighting passages in color. Because the text layer is in the downloaded file, a colleague who opens it in Acrobat or Preview gets the same search without running anything.

What OCR does not do is turn the scan into an editable document. The words are recognized and positioned, but the page is still an image. To change them you need a converter or an editor afterwards; the edit text in a PDF page explains the options.

Troubleshooting OCR results

• Gibberish or wrong letters. The language was wrong. The viewer picks the OCR language from the browser language; switch the browser to the document's language, reload and run OCR again.
• Some words are missed. Low resolution (under about 200 dpi), skewed pages, handwriting and colored backgrounds all lower accuracy. Rescan at 300 dpi in black and white if you can.
• It is taking forever. OCR speed depends on page count and device. Phones and older tablets are much slower than a laptop; for a long scan use a computer.
• The page is sideways. Rotate it from the thumbnail first (see rotate a PDF) so the text runs the right way.
• Search still finds nothing in another app. Make sure you opened the downloaded copy, not the original scan.

Frequently asked questions