DocomintOCR d'un PDF ou d'une image (extraction de texte)
Extrayez le texte d'un PDF scanné, d'une photo ou d'une capture d'écran, entièrement dans votre navigateur.
Drop your PDF or image(s) here
or choose a file from your device
What is a OCR d'un PDF ou d'une image (extraction de texte)?
OCR turns a scanned document or a photo of text — a receipt, a printed article, a screenshot of a book page — into text you can actually copy, search, or paste elsewhere, since a scan is just pixels to a computer until OCR recognizes the characters in it.
How to use the OCR d'un PDF ou d'une image (extraction de texte)
- Add one or more PDFs or images and pick their language
- Each page or image is recognized on your device (a one-time language download the first time)
- Copy the extracted text, or download it as a .txt file
Example
A photo of a printed restaurant receipt run through English OCR returns the line items and total as plain text, ready to paste into an expense spreadsheet instead of retyped by hand.
Why use OCR d'un PDF ou d'une image (extraction de texte)
Works on PDFs and images
Run OCR on a scanned PDF, a phone photo, or a screenshot — the same tool handles all three.
40+ languages
Pick the document's actual language for accurate recognition, instead of one English-only model.
Nothing uploaded
Recognition runs on-device via WebAssembly — your document or photo is never sent to a server.
What can you create?
Digitize a paper receipt
Turn a photographed receipt into text you can paste straight into an expense spreadsheet.
Make an old scan searchable
Extract the text trapped in a scanned report or book page so you can copy quotes out of it.
Pull text from a screenshot
Recover text from a screenshot of an article or slide without retyping it by hand.
Frequently asked questions
Is my PDF uploaded anywhere?
No — OCR d'un PDF ou d'une image (extraction de texte) runs entirely in your browser using JavaScript/WebAssembly. Your PDF is never sent to a server.
Is there a file size limit?
Free use covers files up to 100MB per file; signing up for Pro raises that to 500MB.
How accurate is the extracted text?
Accuracy depends heavily on image quality and language — a clean, high-resolution scan of printed text is typically very accurate, while handwriting, low-resolution photos, or unusual fonts produce more errors that need manual review.
Does this work on handwritten text?
It's built and tuned for printed text; handwriting recognition is a different, much harder problem, and results on handwritten pages will be unreliable.
Can I fix mistakes in the recognized text before copying it?
Yes — the extracted text appears in an editable text area, so you can correct any misread characters before copying or downloading it.
Can I download the extracted text instead of just copying it?
Yes — click "Download .txt" below the results to save the recognized text as a plain-text file.