How to OCR a scanned PDF and make it searchable
Open a scanned PDF and try to select a sentence. If the whole page highlights as one picture, you are looking at a photograph of text, not text. OCR — optical character recognition — reads that photograph and rebuilds a real text layer underneath it, so the document becomes searchable, selectable and copyable, while still looking exactly like the original scan.
First, check whether you need OCR at all
Many PDFs that look scanned are actually born-digital: their text already selects and searches fine, and OCR would add nothing. The five-second test is the selection test above, or searching the file (Ctrl+F) for a word you can see on the page. If search finds it, skip OCR. If the document is a true scan — from a scanner, a camera app, or an old fax — OCR is the unlock for everything else: after OCR you can also convert the file to Word or Excel with far better results.
How browser OCR works
Our OCR PDF tool runs a recognition engine (Tesseract) entirely in your browser — unusual, because most OCR services upload your documents to their servers. Each page is rendered, the engine identifies letter shapes and assembles words, and the recognised text is embedded back into a copy of the PDF as an invisible layer aligned with the image. Choosing the right document language matters: it tells the engine which alphabet and dictionary to expect, and it is the single biggest accuracy setting you control.
Step by step
1. Open the OCR PDF tool and load the scan
Everything happens on your device, so even a long document is private — but a long document also takes a while; OCR is real work, page by page.
2. Set the language
Pick the language of the document (or the closest match for mixed pages). Wrong language is the most common cause of gibberish output.
3. Run it, then test the result
Download the searchable copy and try the tests from the first section: select a sentence, search for a word from the middle of the document. Spot-check names and numbers against the scan — those are what OCR gets wrong first.
Scan habits that decide accuracy
- Resolution: 300 DPI is the sweet spot. At 150 DPI small letters merge; much above 300 only adds time.
- Straight pages: a few degrees of tilt is fine; a page photographed at an angle confuses line detection. Rotate first if needed.
- Contrast and light: grey shadows across the page read as dirt. Scan in good light, in colour or greyscale rather than pure black-and-white mode.
- Handwriting is the honest limit: printed text recognises well, handwriting mostly does not, whatever the tool promises.
What a searchable PDF unlocks
The OCR result is not the end of the job; it is the beginning of several others. Finding things: a searchable archive means Ctrl+F replaces re-reading — search a name, a date, an amount across a contract or a decade of statements. Quoting and copying: select the paragraph and paste it into an email or a report instead of retyping it (proofread the paste: recognition of names and numbers is where errors concentrate). Conversion: with a text layer in place, the same file converts to Word for editing or to Excel for its tables, with results that were impossible while it was a photograph. Translation and summarising: text-based tools — our Translate PDF and the local summarizer — can only work on documents whose words exist as text. Keep both versions: the original scan as the visual record, and the OCR copy as the working document, named so you never confuse them.
Frequently asked questions
Will OCR change how my document looks?
No. The pages keep their scanned appearance; OCR adds an invisible text layer underneath so you can search and select.
How accurate is OCR?
On a clean, straight 300 DPI scan of printed text, expect near-perfect results. Accuracy drops with tilt, shadows, small fonts and handwriting — always spot-check names and figures.
Are my files uploaded for OCR?
No. The recognition runs 100% in your browser; your document never leaves your device — which is precisely why it may take a little longer on big files.
Open the free OCR PDF tool →Runs 100% in your browser — your file is never uploaded.