OCR PDF
Free OCR for PDFs: add a hidden text layer to a scanned PDF so you can search and copy it. Looks identical, runs in your browser, no upload.
The 10-second test: switch off your Wi-Fi (or turn on airplane mode) and use this tool anyway. It keeps working — because there is no server doing the work.
The technical check: open your browser's developer tools, go to the Network tab, and run the tool. Your file is never sent anywhere.
Measured live on this page — files transmitted: 0 · file data sent: 0 bytes
This counter hooks the browser's own networking functions, so it would rise the instant any file data left this page. The only thing this site ever transmits is the contact form, if you choose to send one.
Select PDF files
or drag & drop — pasting works too (Ctrl+V)
Working…
Your file is ready!
Processed on your device — 0 bytes uploaded
Thanks — feedback received!
How to make a scanned PDF searchable
A scanned PDF is really just photos of paper — Ctrl+F finds nothing because there is no text inside. OCR fixes that: we read the words on each page and tuck an invisible text layer behind the image. The document looks exactly the same, but now it is searchable, selectable and copyable.
This is the feature Adobe Acrobat Pro calls "Recognise Text" and charges a subscription for. Here it is free, and the scan never leaves your device — which is the point, because the documents people scan are contracts, statements and records.
What OCR is for
A scanned PDF looks like a document but behaves like a photograph. You cannot select the text, search it, or copy a sentence out of it, because as far as the file is concerned every page is one big image. OCR reads the shapes of the letters and writes a real, invisible text layer underneath, so the page looks identical but becomes searchable and selectable.
What affects accuracy
Resolution first: 300 DPI is the sweet spot, and anything under 200 DPI degrades quickly. Straightness second — a page scanned at an angle loses accuracy fast, and a photo taken at a slant is worse. Contrast third: faint photocopies and yellowed paper are hard. Printed text in a normal font is read very reliably; handwriting is not supported at all, and unusual display fonts are unreliable.
Why it takes a while
The recognition engine is roughly 15 MB and downloads once, then works offline. Recognition itself runs on your own processor, which is why a long document takes real time — but also why your document is never uploaded. Medical records, contracts and financial statements are exactly the documents people OCR, and exactly the ones that should not go to a stranger's server.
After OCR
Once the text layer exists you can search it in any reader, copy text out of it, or convert it with Scanned PDF to Word or PDF to Text.
Frequently asked questions
Will my PDF look different afterwards?
No. The original pages are untouched — the text layer is invisible and sits behind them. Only the file size grows slightly.
How do I check it worked?
Open the downloaded PDF and press Ctrl+F (Cmd+F on Mac), then search for a word you can see on the page. You can also select text with your cursor.
How long does it take?
Roughly 2-8 seconds per page, depending on the page. The OCR engine (~15 MB) downloads once on first use, then works offline.
What DPI should I scan at for OCR?
300 DPI is ideal. Below 200 DPI accuracy drops sharply; above 400 DPI you gain little and the file gets much larger.
Does OCR work on handwriting?
No. Printed text in ordinary fonts is read reliably; handwriting is not supported.
Why does OCR take so long?
Recognition runs on your own processor rather than a server. That is slower than a data centre, but it is the reason your document is never uploaded.
Does the page look different afterwards?
No. The scanned image is unchanged; an invisible text layer is added beneath it.