Skip to main content

Scanned PDF to Text

Scanned PDF to Text

Turn scanned paper documents into text you can search, correct and export.

Read text from scanned PDFs

Drop photocopied or scanned PDFs. Recognition runs in your browser.

Max 100MB per fileUp to 50 files

How to OCR Scanned PDFs to Editable Text

How to Use Scanned PDF to Text

  1. 1

    Upload Scanned PDF

    Select one or more scanned PDF files.

  2. 2

    Select Recognition Language

    Pick one of 55 languages (English, Spanish, German, French, Hindi, Japanese and more), and a second one for a document that mixes two.

  3. 3

    Adjust Preprocessing If Needed

    Black-and-white conversion is on by default. Brightness, contrast, the threshold and the deskew angle are yours to set — nothing is estimated for you.

  4. 4

    Run OCR & Download

    Correct anything the engine misread, then save as TXT, DOC, JSON, a table export or a searchable PDF.

What Scanned PDF to Text Can Do

  • In-Browser Tesseract OCR

    Your pages are rasterised and recognised on your own device — no image data is sent anywhere.

  • Preprocessing You Control

    Brightness, contrast, a black-and-white threshold and a deskew angle, previewed live on the page before you run.

  • 55 Recognition Languages

    Latin, Cyrillic, Greek, Devanagari, Arabic, CJK and more, with an optional second language for documents that mix two.

  • Searchable PDF Output

    Saves the original pages with the recognised words laid invisibly on top, so the PDF becomes searchable. Latin, Greek and Cyrillic words go into that layer; the tool tells you when words in other scripts cannot.

Tools & Settings Reference

OCR Language Selection
One of 55 languages, with an optional second one for a document that mixes two. The engine and each language are downloaded from a public CDN the first time they are used, so that step needs a connection.
Recognition Mode
Smart reads a page that already carries real text directly — faster and exact — and only OCRs the rest. Force OCR rasterises every page.
Binarize & Threshold
Converts the page to pure black and white at the grey level you choose. It is a fixed cut-off, not an automatic one, and there is no noise removal.
Deskew Rotation
Straightens a crooked scan by the angle you set before it is recognised. There is no automatic skew detection, and in Smart mode a page read from its own text layer is not affected.
Low-Confidence Threshold
Recognised words scoring below this are underlined red in the Verification pane so you can check them.

Frequently Asked Questions

The page is drawn at 300 DPI (a little less for a page too large for the browser to draw at once) and then passed through the filters you set: a brightness offset, a contrast curve, and a black-and-white conversion at a grey threshold you choose (128 by default). A deskew angle you set rotates the page before recognition. All of it is manual — there is no automatic threshold estimation, no skew detection and no despeckling — so a poor scan is worth a couple of passes with the sliders and the live preview.

Yes. Pick the main language and add the other as the second language, and both are used together on every page, so a Hindi letter with English terms comes out in both scripts. Each language is a separate download and recognition gets slower with two, so leave the second one empty when you do not need it. Most Latin-script languages already handle the odd English word on their own.

Your pages do not leave the browser — the recognition itself is local. The OCR engine, its WebAssembly core and the language pack you pick are downloaded from a public CDN the first time they are needed, so the first run needs an internet connection and the language you chose is visible to that CDN. Nothing about the document itself is sent.

Related Guides

Related Tools