OCR PDF – Extract Text from Scans
Read the text in scanned PDFs and photos and save it as a text file.
🔒 Your file is processed inside your browser. It is never uploaded or saved.
What is OCR PDF?
A scanned PDF looks like text but is really a picture, so you cannot search or copy it. Optical character recognition reads the letters in each page image and turns them into real text that you can copy, search and edit. This tool supports PDFs and common image files and offers nine recognition languages.
How to OCR a scanned PDF
- Add a scanned PDF or a photo of a document (JPG, PNG, WebP or BMP).
- Choose the language of the text in the Language list so recognition matches the script.
- Press Start. The first run downloads the open-source recognition engine, then each page is read in turn.
- Download the .txt file, which separates pages with a marker so you can find your way around.
When OCR PDF is useful
Searching old scans
Make the words of a scanned letter or book chapter available for copying.
Editing a scanned form
Get the wording into a word processor so you can rewrite it.
Capturing a photo of text
Pull a paragraph from a phone picture of a poster or page.
Tips for better results
Accuracy depends on the scan. Straight, well-lit pages at a decent resolution give the best results, while blur, shadows and handwriting reduce it. Always proofread the output. The recognition engine and language data are downloaded from public content networks when you use this tool, so you need an internet connection, but your document itself is not sent anywhere.
Why use PDFToolkit
When you OCR a scanned PDF with PDFToolkit, the work stays on your own device. Your contracts, ID scans and invoices are processed in the browser and are never sent to a server.
- Private: no upload, nothing saved.
- Free: no sign-up, no daily limit and no watermark.
- Fast: no waiting for files to travel over the internet.
- Everywhere: Windows, Mac, Linux, Android and iPhone.
Frequently asked questions
Which languages are supported?
English, Urdu, Hindi, Arabic, Spanish, French, German, Russian and Simplified Chinese.
Is my file uploaded for OCR?
No. Only the recognition engine files are downloaded to your browser. Your document stays on your device.
What output do I get?
A plain text (.txt) file with the recognised text, divided by page.
Why are some words wrong?
Low-resolution or skewed scans, unusual fonts and handwriting make recognition less accurate, so check the text.