Runs in your browser · nothing is uploaded

Make a scan searchable.

The words go back into the file. The page looks exactly as it did.

PDF OCR

Drop a scanned PDF here

or pick one from your device

Everything happens on your device

Reading…

Is your PDF a scan?

Ctrl+F finds nothing
You can see the word on the page and the search box says it is not there.
You cannot select a line
Dragging over the text draws a box around the whole page instead of highlighting words.
It came from a scanner or a phone
Anything photographed or scanned is pixels. So is most of what a fax or a copier produces.

50 MB per file · English

Checking which pages have text…

What is in this PDF

Locked document

This PDF needs its password

A few seconds a page. Keep this tab open.

Done

By using this tool you agree to our terms and privacy policy

What it actually does

The words go underneath

Each scanned page is rendered at about 190 DPI, read, and the recognised words are written back into the file in the PDF rendering mode that draws nothing — placed exactly over the picture of each word.

The page looks identical
Nothing is painted over your scan and nothing is re-encoded. The image you uploaded is the image in the file you get back — the only difference is a layer of text you cannot see.
It works in every reader
Because the text is in the file rather than in this tab, Preview, Acrobat, your phone, your document management system and anything that indexes PDFs can all search it. Selecting and copying a paragraph works too.
Pages that have text are left alone
A document that is half typed and half scanned is common, and re-reading the typed half would replace text that is exactly right with text that is a good guess. Only the scanned pages are touched, and the map above says which.
OCR guesses, and says so
Every word carries a confidence score, and the ones that were hard to read are counted on screen. Check anything that matters — a 1 read as an l in an account number is the failure to look for.

Answers

Questions people ask

Does my file get uploaded?

No. The recognition engine is downloaded into your browser the first time you run OCR, and the pages are read there. Nothing about your document is sent anywhere — we record only an anonymous count that a run happened.

Why is the first page slow?

The engine and its English language data are about 8 MB, downloaded once when you press the button. After that each page is a few seconds. A long document is genuinely long — this is real work happening on your own machine rather than a queue on a server.

Which languages does it read?

English. The recognition model shipped with this page is the English one, and a French or Spanish page will come back with plausible-looking mistakes rather than an error, which is why it is worth saying plainly here.

Can I edit the text afterwards?

You can search, select and copy it anywhere. Changing what the page says is a different job — open the file in the PDF editor, which can also run this same OCR and then let you edit lines.

Can I black out something on a scan?

Not safely here, and it is worth knowing why. On a scan the words are pixels inside an image, so a black box drawn over them hides them from your eyes without removing anything from the file. This tool will never write recognised text into an area you have blacked out — but making a scan genuinely safe to send means removing the pixels, which the editor's redaction tool reports honestly on a page it cannot do it for.

Make a scan searchable

Free, no account, no watermark, and the file never leaves your device.