A scanner produces a picture of paper. The words look right to you, but Ctrl+F finds nothing and copy-paste returns nothing. OCR (optical character recognition) adds a real text layer under the image so the document becomes searchable without changing what it looks like.
Step by step
-
1
Upload your PDF
Drag your PDF onto the box or click Choose file. It is sent over an encrypted connection and only used for this task.
-
2
Set the options
Adjust language and ocr engine. Every field has a sensible default, so you can usually run the tool straight away.
-
3
Runs in your browser
OCR uses tesseract.js on your own machine, so a scanned document never has to be uploaded. It adds a text layer, which makes the scan searchable and selectable — the page image is untouched.
-
4
Run it and download
Press OCR and the page shows progress while it works. When it finishes you get a download link — nothing is emailed and nothing is stored longer than it needs to be.
What OCR changes — and what it leaves alone
Recognition reads the shapes on the page and writes a text layer behind the image. The scan still looks the same, but the words become selectable and searchable, and screen readers can announce them. It does not magically repair a blurry page or turn a photograph of a handwritten note into perfect text.
Prepare the scan before recognition
Accuracy is mostly decided before OCR starts. Scan at roughly 300 DPI, keep the page flat and square, and straighten crooked pages with Deskew first. A dark background or uneven phone lighting can confuse recognition; a clean, evenly lit page is worth more than any post-processing setting.
Language, layout and handwriting
Choose the language that appears on the page. Multi-column layouts are read in an order that depends on the document, so check the output rather than assuming it follows the visual columns. Handwriting is the hardest case: OCR may recognise printed labels around it but will often miss the handwriting itself. Treat that as a limitation, not a bug.
Check before you rely on the text
Search for a few distinctive words and numbers — an invoice total, a clause number, a surname — and compare the result with the page. Open the text layer in a viewer and check that the page still prints correctly. The OCR engine runs in your browser on this site, so the scan does not need to be uploaded for recognition; that privacy difference matters for passports, medical records and client files.
Good habits for OCR
- OCR runs in your browser, so scanned documents are never uploaded.
- Straighten the scan first — deskewed pages give noticeably better OCR.
- A 300 DPI scan is the sweet spot between accuracy and speed.
Common questions
Is this tool free?
Yes. Every tool on this site is free to use, with no sign-up, no watermark and no daily task limit.
Do I need to install anything?
No. The tools run in your browser and are processed by this server.
What happens to my files?
They are stored under an unguessable id and swept automatically after a couple of hours. Nothing is shared or indexed.
Does it work on a phone?
Yes. The page pickers and editors are built for touch, so dragging, rotating and selecting pages all work without a mouse.
What does OCR do exactly?
Scanned pages get a searchable, copyable text layer.
Does OCR change the scanned image?
No. It adds a text layer behind the image. The page still looks like the original scan.
Does OCR work on handwriting?
Sometimes, but it is unreliable. Printed text is the strong case; handwritten notes often need manual transcription or correction.
Is the scan uploaded for OCR?
On this site recognition runs in your browser with Tesseract.js. The searchable PDF is assembled locally, so the scan itself is not sent away for recognition.