OCR & Documents
Convert a scanned document or photo to Word
Free to useNo sign-up requiredNo watermarkRuns in your browser
A scanned PDF or a photo of a page has no real text in it - OCR to Word reads it, recognizes the words, and rebuilds them as an ordinary Word document you can edit, rather than leaving you to retype the whole thing.
This is the same recognition engine as the OCR tool, taken one step further: instead of stopping at plain text, it reconstructs paragraphs, headings and tables so the result reads like a document rather than a wall of text.
How this tool works
Upload your scan or photo
JPG, PNG, WebP or a scanned PDF, read directly in your browser.
Text is recognized automatically
A progress indicator shows what is happening, page by page for a multi-page PDF.
Choose your options
Rebuild tables as real Word tables, and add a heading before each page.
Download the .docx
Opens in Word, Google Docs, LibreOffice and Pages.
How it works
Recognition runs on the same self-hosted Tesseract engine as the OCR tool. For a scanned PDF, each page that has no real text is rendered as an image and OCR is run on it; a page that already has real text (a mixed PDF) is read directly instead, keeping its actual layout.
The recognized words, each with their position on the page, are grouped into rows and columns the same way PDF to Word groups a real PDF's text - clustering positions rather than assuming a fixed layout - so a detected table becomes a real Word table, not a row of text with big gaps in it.
Nothing about this is a picture of the page turned into a Word file: the output is real, selectable, editable text, built from what OCR recognized.
What to expect
This is a recognition, not a scan-to-file conversion - accuracy depends entirely on how clear the source is. A clean, straight scan or a well-lit, in-focus photo converts well; a blurry, tilted or low-resolution one will have mistakes worth checking.
- Clear typed or printed text: reliable.
- Handwriting: not reliable - this engine is built for printed text.
- Rotated or sideways pages: rotate the page first with Rotate PDF, then convert.
- A PDF that already has real, selectable text: use PDF to Word instead - it keeps the original fonts and layout information OCR cannot see.
Worked examples
A photographed contract page
Upload the photo and download a Word document with the text recognized and laid out as paragraphs, ready to edit.
A scanned report with a table
The table is detected from the recognized text's layout and rebuilt as a real, editable Word table rather than misaligned text.
Frequently asked questions
Is my file uploaded to a server?
No. Recognition and the Word document are both built entirely in your browser, on your own device.
What is the difference between this and PDF to Word?
PDF to Word reads a PDF's own text layer, which already knows the exact fonts and positions. OCR to Word is for when there is no text layer at all - a scan or a photo - and has to recognize the words first.
Can it read handwriting?
Not reliably. This engine is built and trained for printed and typed text, not handwriting.
Is there a file size or page limit?
Yes - OCR runs on your own device rather than a server, so very large files and very long PDFs are refused with a clear explanation rather than being allowed to freeze the tab.