All tools
All tools › PDF OCR - Scanned PDF to Text ● Live PDF Tools

How PDF OCR - Scanned PDF to Text works

  1. Select the language the document is written in — this strongly affects accuracy.
  2. Drop a scanned PDF onto the drop zone, or click it to browse.
  3. Watch per-page progress as each page is rendered and recognized — a few seconds per page is normal.
  4. Review the extracted text (pages are separated by '--- Page N ---' markers), fix any mistakes, then copy it or download it as a .txt file.
  • Turn a scanned contract, report, or letter into text you can quote and search.
  • Recover the text of an old printed document that only exists as a scan.
  • Extract text from a photographed document saved as PDF by a phone scanning app.
  • Make the content of an image-only PDF quotable without retyping it.
  • Pull text out of a fax or photocopy archive saved as PDF.

The PDF is opened, rendered, and recognized entirely on your device. It is never uploaded to a server. Only the OCR engine and language data files are downloaded, and they are cached for later runs.

  • OCR is probabilistic: even clean scans can contain recognition errors, so always proofread the output.
  • Limited to 50 pages per run — split longer PDFs first.
  • Password-protected or corrupted PDFs cannot be processed.
  • Handwriting, cursive text, and complex multi-column layouts are recognized poorly or not at all.
  • Output is plain text only — formatting, tables, and images are not preserved.
  • The engine and language data (a few megabytes) are downloaded on first use, so the first run needs a network connection.

PDF OCR - Scanned PDF to Text

PDF OCR turns scanned PDFs into editable text without uploading the file anywhere. Each page is rendered in your browser and recognized by the Tesseract OCR engine compiled to WebAssembly. It complements the PDF Text Extractor: use that tool for PDFs with embedded text, and this one for scans and image-only PDFs where there is no text layer to extract.

PDF OCR - Scanned PDF to Text Runs locally

Pick the language the document is written in — it strongly affects accuracy.

Drop a scanned PDF here, or click to browse

Each page is rendered and recognized locally — up to 50 pages
Similar tools

More PDF Tools tools

See all tools →
LivePDF Tools

Searchable PDF Converter (OCR)

Make a scanned PDF selectable and searchable by adding an invisible OCR text layer.

searchable pdfocr pdfselectable text
LivePDF Tools

Image to PDF

Convert one or more images into a PDF file in your browser.

imagePDFconverter
LivePDF Tools

PDF Page Splitter

Split a PDF into individual pages or extract a page range.

PDFsplitextract pages
LivePDF Tools

PDF Text Extractor

Extract all text from a PDF file directly in your browser.

PDFtext extractionlocal
LivePDF Tools

PDF to Image

Convert PDF pages to PNG images locally in your browser.

PDFimagePNG
LiveImage / Inspection

Image to Text (OCR)

Extract text from images with OCR that runs entirely in your browser.

ocrimage to textextract text
Guide

How to use

  1. Select the language the document is written in — this strongly affects accuracy.

  2. Drop a scanned PDF onto the drop zone, or click it to browse.

  3. Watch per-page progress as each page is rendered and recognized — a few seconds per page is normal.

  4. Review the extracted text (pages are separated by '--- Page N ---' markers), fix any mistakes, then copy it or download it as a .txt file.

Scenarios

Use cases

  • Turn a scanned contract, report, or letter into text you can quote and search.

  • Recover the text of an old printed document that only exists as a scan.

  • Extract text from a photographed document saved as PDF by a phone scanning app.

  • Make the content of an image-only PDF quotable without retyping it.

  • Pull text out of a fax or photocopy archive saved as PDF.

Good to know

Limitations & Privacy

Private by design

The PDF is opened, rendered, and recognized entirely on your device. It is never uploaded to a server. Only the OCR engine and language data files are downloaded, and they are cached for later runs.

  • OCR is probabilistic: even clean scans can contain recognition errors, so always proofread the output.

  • Limited to 50 pages per run — split longer PDFs first.

  • Password-protected or corrupted PDFs cannot be processed.

  • Handwriting, cursive text, and complex multi-column layouts are recognized poorly or not at all.

  • Output is plain text only — formatting, tables, and images are not preserved.

  • The engine and language data (a few megabytes) are downloaded on first use, so the first run needs a network connection.

FAQ

Frequently asked questions

Is PDF OCR free to use?

Yes. The tool is free and runs directly in your browser with no account required.

Does this tool upload my PDF?

No. Each page is rendered to an image and recognized by the Tesseract OCR engine compiled to WebAssembly, all inside your browser tab. Only the engine and language data are downloaded, once, and then cached.

How is this different from the PDF Text Extractor?

The PDF Text Extractor reads the text layer embedded in digitally created PDFs — fast and exact. Scanned PDFs have no text layer, only page images, so extraction returns nothing. PDF OCR recognizes the text from those images instead. If you're not sure which you have, try the extractor first; if it comes back empty, your PDF is a scan.

How many pages can it handle?

Up to 50 pages per run, to keep memory use reasonable. For longer documents, split the PDF first (the PDF Split tool on this site works) and run each part.

Why is it slow on my document?

OCR is computationally heavy: each page is rendered at high resolution and analyzed pixel by pixel on your own device. A few seconds per page is normal; older devices take longer.

Can it handle a password-protected PDF?

No. Password-protected and encrypted PDFs cannot be opened. Remove the password first with a PDF tool that knows the password.

Why is my result garbled or full of mistakes?

Usual causes: the wrong recognition language selected, a faint or skewed scan, handwriting (not supported), or complex multi-column layouts. A clean, straight, 300-DPI-equivalent scan in the selected language gives the best results.

Does it preserve the document's formatting?

No — the output is plain text with a '--- Page N ---' marker between pages. Columns, tables, fonts, and images are not reconstructed.