How to OCR a Scanned PDF
Updated
A scanned PDF is just pictures of pages, so you can't search it or copy its text. OCR (optical character recognition) reads the words and adds them as an invisible text layer. Here is how to do it privately in your browser.
Step 1: Add the scanned PDF
Open OCR PDF and select your scan, or several. The text recognition engine runs inside your browser, so the document never leaves your device.
Step 2: Choose the language
Pick the language the document is written in: English, Hindi, Bengali, Tamil, Telugu, Marathi, Gujarati, Kannada, Malayalam, Punjabi, Odia, Assamese, Urdu, Sanskrit or Nepali. For Indian-language documents, English words such as names and numbers are read too unless you switch that off. Choosing the right language matters more than anything else for accuracy.
Step 3: Choose the output
A searchable PDF looks the same as before, but the text can now be selected, copied and found with Ctrl+F. A text file gives you just the words, ready to paste into another document.
Step 4: Process
Click "Process Files". Each page takes a few seconds. The first run also downloads the recognition engine and the language file, about 4 to 10 MB, from this site.
Crooked and sideways pages
Pages scanned at a slight tilt, up to about 12 degrees, are straightened automatically; you can switch this off. Pages photographed at a steep angle should go through Scan to PDF first, where you can mark the corners and flatten the page.
Pages that already have text
By default these are skipped so the text isn't doubled. Choose "OCR every page" if the existing text is wrong or garbled.
Tips for better results
Use clear, well-lit scans of about 200 to 300 DPI. Very small print, blurry photos and handwriting are read poorly. Each PDF can be up to 50 pages; split longer documents with Split PDF first.
Afterwards
To copy out all the text from a PDF that already has a text layer, use Extract PDF Text. OCR adds very little to the file size, but if the scan itself is large, Compress PDF will shrink it.