OCR PDF

OCR a scanned PDF

Turn scanned or image-based PDFs into searchable, selectable text with optical character recognition.

OCR a scanned PDF online

A scanned PDF is a stack of pictures: you can read it, but Ctrl+F finds nothing. PDF Vision runs optical character recognition on each page and writes the recognised words behind the page as an invisible text layer, so the document looks the same while becoming searchable, selectable and copyable. Recognition is forced on every page, which re-renders and re-encodes it, so use this on scans rather than on PDFs that already carry good text. Pick the main language of the scan — English, French, German, Spanish, Italian, Portuguese, Russian, Arabic, Chinese or Japanese — then download the searchable PDF. Processing runs on our servers and OCR is free with a free account.

How to oCR a scanned PDF

  1. Sign in and upload

    Open your free PDF Vision account, then select the scanned or image-based PDF you want to make searchable.

  2. Choose the scan language

    Pick the language the document is printed in; the engine transcribes characters as written and never translates them.

  3. Download the searchable PDF

    Long scans take a few minutes to process, then the finished PDF downloads with its text layer already embedded.

Why use it

Ctrl+F works again

The page is re-rendered as it looks today and the recognised words are written behind it, so stamps, signatures and layout stay visually identical, while search, copy and text selection behave like a born-digital PDF. Because OCR is forced on every page, run it on scans, not on files that already have a good text layer.

Ten recognition languages

English, French, German, Spanish, Italian, Portuguese, Russian, Arabic, Simplified Chinese and Japanese are covered, including non-Latin scripts that naive text extraction usually returns as unreadable characters.

Unlocks downstream tools

Once a text layer exists, the same file can be converted to Word or Excel, translated, summarised or redacted — none of which work on a page that is only an image.

Frequently asked questions

What does OCR actually change in my PDF?
Visually almost nothing: stamps, signatures and layout look the same, and a hidden text layer is added underneath. Technically, OCR is forced on every page, so each page is re-rendered and re-encoded, the file size changes, and any selectable text the PDF already had is replaced by OCR text — which may be less exact than the original.
Do I need an account to run OCR?
Yes. OCR PDF is free but requires a free PDF Vision account. Recognition runs on our servers and takes real processing time on multi-page scans, so it is tied to a signed-in session.
Which languages can it recognise?
English, French, German, Spanish, Italian, Portuguese, Russian, Arabic, Simplified Chinese and Japanese. English pairs with French, and French, German and Spanish each pair with English, so a bilingual page is read in a single pass.
Will OCR translate my document?
No. Recognition transcribes the characters exactly as printed on the page. If you also need the content in another language, run Translate PDF on the searchable file once OCR has finished.
How accurate is the recognition?
Accuracy follows scan quality. A straight, high-contrast 300 DPI scan is read almost perfectly. Faxes, skewed pages, handwriting, stamps over text and heavy background noise produce errors you should proofread before relying on the text.
How long does a long scan take?
Processing time grows with page count, and a large scan can take several minutes — the request times out at ten. Keep the tab open until the download starts rather than reloading the page.

Related tools