SumoPDF

OCR a PDF (make it searchable)

SumoPDF mascot

Recognize the text in a scanned or image-based PDF and add an invisible, searchable text layer, entirely in your browser. Nothing is uploaded.

Drop one PDF here or browse

How it works

  1. Add a scanned or image-based PDF.
  2. Choose the document's language.
  3. Select Make searchable and review. SumoPDF reads each page and adds a selectable text layer over the original, so you can search and copy the text. The look of the pages is unchanged.

Limits

Browser-local, powered by an in-browser OCR engine (a one-time engine and language-data download of a few megabytes). It is slower than the other tools and works best on clear scans; handwriting and low-quality images may not read well. English and Spanish. OCR gives 2 free uses a day for documents up to 25 pages. Upgrade to SumoPDF Pro for unlimited uses and no page limit. Password-protected PDFs are not supported. This tool runs a WebAssembly OCR engine, so the OCR page enables WebAssembly in its security policy (unlike the other tools); your file is still processed on your device and never uploaded.

Common questions

Is my document's content secure when I use OCR online?

Yes, your document's content is kept secure. The OCR process runs in your browser, meaning the file and its recognized text are never uploaded to a server. Your data remains private on your device.

What does it mean to OCR a PDF?

OCR stands for Optical Character Recognition. It is a process that analyzes an image of text, like a scanned document, and converts it into actual text characters. This makes the PDF's content searchable and allows you to copy and paste the text.

Can I use OCR on a long, scanned book?

OCR gives 2 free uses a day for documents up to 25 pages. To process a longer scanned book or get more uses, upgrade to SumoPDF Pro for unlimited uses and no page limit.

In which languages can you recognize text?

The OCR tool is optimized for recognizing text in English and other Latin-script languages. The accuracy of the text recognition can vary depending on the quality and clarity of the original scan.