PDF OCR – Extract Text from Scanned Documents

How to extract text from a scanned PDF

1. Upload your scanned PDF by clicking Choose File or dragging it into the drop zone.
2. Select the document language from the dropdown — 41 languages supported.
3. Click Run OCR — VoxScan extracts all readable text and lets you download it as a .txt file.

When to use PDF OCR

Scanned documents look like PDFs but contain only images — you can't search, select, or copy the text. OCR (Optical Character Recognition) reads those images and converts them into real, searchable text. Use it to digitize paper contracts, make scanned invoices searchable, extract data from bank statements, or convert printed research into editable content. VoxScan supports 41 languages including Arabic, Chinese, Japanese, Korean, Hindi, and all major European languages.

From OCR to structured XML

OCR gives you the text; sometimes what you actually need is structured data. A scanned invoice, delivery note, or form is only useful to a script or an accounting system once its contents exist as machine-readable fields, and the path there is a two-step workflow: first run the scan through OCR on this page so the document gains a real text layer, then feed the result to the PDF to XML converter, which serializes every text run into XML elements with page numbers and x/y coordinates. The coordinates matter — they let downstream code group lines into table rows, pull a total from a known region of the page, or reconstruct reading order, none of which is possible with a flat .txt export. Both steps run entirely in your browser, so a confidential contract or bank statement never touches a server on its way from paper scan to structured XML. If the scan is faint or low-contrast, run it through the brightness & contrast editor first: OCR accuracy climbs sharply once the text is dark and the background clean, and so does the quality of the XML built on top of it. For a worked end-to-end example, see the guide on converting a PDF invoice to XML.

Frequently asked questions

Which languages does the OCR support?

41 languages: English, Spanish, French, German, Italian, Portuguese, Dutch, Polish, Russian, Turkish, Arabic, Chinese (Simplified & Traditional), Japanese, Korean, Hindi, and more.

How accurate is the OCR?

Accuracy depends on scan quality. Clean, high-contrast scans at 200 DPI or above typically achieve 95%+ accuracy. Low-resolution or skewed scans may produce errors that need manual correction.

Can OCR handle handwritten text?

VoxScan's OCR is optimised for printed text. Handwriting recognition is not currently supported.

Can I convert a scanned PDF to XML?

Yes, in two steps. Run the scan through OCR here so the document gains a searchable text layer, then open the result in the PDF to XML converter on this site. The output is structured XML where each text element carries its page number and position coordinates, ready for parsing, data extraction, or import into another system.

Does my scanned document leave my device during OCR?

No. The OCR engine runs inside your browser via WebAssembly. Pages are rendered and recognized locally, and the extracted text is assembled on your device — no image data or text is uploaded to VoxScan or any third party, which matters for contracts, medical records, and financial documents.

Is the OCR really free, and is there a page limit?

Yes — no page cap, no daily quota, no watermark, and no account required. Because recognition happens on your own hardware, the practical limit is your device's memory and patience: long documents simply take longer to process, roughly a few seconds per page on a typical laptop.