MarkdownPDF

OCR a PDF — make scans searchable

Turn a scanned PDF into one you can select, search and copy from. The pages look exactly the same — the recognised text sits invisibly on top. Seven languages, free, and nothing leaves your browser.

Drop a scanned PDF, or browse

Processed on your device — never uploaded

A scan is a picture, not a document

When a page comes off a scanner, the PDF holds an image of it. To your eyes it is a document; to every piece of software it is a photograph. Ctrl+F finds nothing, no sentence can be selected, and an archive of a thousand such files is unsearchable.

OCR closes that gap. It reads the picture, recognises the characters, and writes them back into the same PDF as an invisible layer aligned with the printed words. Nothing on the page moves; the document simply starts answering searches.

Mixed documents keep their quality

Most real files are mixed: a typed cover page, a scanned contract, a typed appendix. Pages that already contain text are copied from your original untouched — same vectors, same fonts, same file weight — and only the scanned ones are rebuilt. You never lose quality on a page that did not need OCR.

If you want the text out instead

This tool keeps the document and adds text to it. When you want the content itself — to quote, edit or hand to an AI tool — use PDF to Text for the words alone or PDF to Markdown to keep headings and lists as well. Both run the same OCR on the same pages. To get the best out of any of them, the scanned PDF guide explains what actually improves accuracy, in order.

Frequently asked questions

What does OCR do to my PDF?

It looks at each scanned page as an image, recognises the characters, and writes them into the PDF as an invisible text layer sitting exactly over the printed words. The page looks identical, but you can now select, search and copy the text — and so can any other software.

Does it change how the document looks?

No. Pages that already contain text are copied from your original untouched, at full quality. Scanned pages keep the same image with the text layer added on top.

Which languages are supported?

English, French, Spanish, German, Portuguese, Italian, Arabic. Pick the language of the document before converting — recognition uses a language model, so the wrong choice measurably lowers accuracy, especially on accented text.

Searchable PDF or plain text — which do I want?

A searchable PDF when the document must stay a document: contracts, invoices, archives you need to keep as they are but be able to search. Plain text or Markdown when you want the content out of it, to edit, quote or feed to another tool.

Why is the file bigger afterwards?

Recognised pages are rebuilt from the image the engine read, and the text layer adds its own data. The increase is usually modest. Pages that already had text add nothing, because they are copied as they were.

Is anything uploaded?

No. The OCR engine, its language data and the PDF writer all run in your browser, and the engine is served from this site rather than a third-party CDN. Scanned documents are exactly the kind of file that should not be handed to a server.

Limitations

Guides