Image to Word

Processed in your browser. Nothing is uploaded.

There is text locked inside an image, a scanned letter, a screenshot, a photo of a page, and you need it as editable text, not a picture. This tool reads the words in the image and drops them into a Word document you can edit, copy and reformat like any other file. Upload the image, let it recognise the text, download a .docx.

How to convert an image to Word

  1. 1

    Upload an image with text

    A clear screenshot, scan or photo works best. Printed text reads most accurately.

  2. 2

    The text is recognised

    Optical character recognition reads the words out of the picture.

  3. 3

    A Word document is built

    The extracted text is placed into an editable `.docx` you can open in Word.

  4. 4

    Download and edit

    Open it in Word, Google Docs or LibreOffice and change anything you like.

What makes a good source image

Factor Best for accuracy Hurts accuracy
Sharpness Crisp, in focus Blurry or low resolution
Lighting Even, no glare Shadows, reflections
Angle Straight-on, level Tilted or skewed photos
Text Printed, clear font Handwriting, stylised fonts
Background Plain, high contrast Busy or low-contrast backdrops

Where this helps

  • Scanned documents you only have on paper.
  • Screenshots of text you cannot select or copy.
  • Photos of book pages, signs or notices you want to quote or edit.
  • Old printouts with no digital source.

What carries over and what does not

The tool focuses on the words. Plain paragraphs come across well; complex page layouts, columns, tables and exact fonts may not be reproduced precisely. Think of it as a head start on editable text rather than a pixel-perfect copy of the original page.

Formats, limits and privacy

Choose one PNG, JPG or WebP image up to 20 MB, 24 megapixels and 8192 pixels on either side. Select from 18 OCR languages. Recognition, text editing and DOCX or TXT export happen in your browser; the image and text are not uploaded, sent to our servers or put in the URL. In funnel mode they remain in this browser’s IndexedDB for up to 30 minutes, and Start over removes them. The browser still downloads the OCR library and the selected language model.

Always proofread

Optical character recognition is strong but imperfect, lookalike characters, faint print and unusual fonts can produce typos. Read the Word document against the image and clean up any slips. It is still far faster than retyping the whole thing from scratch.

Frequently Asked Questions

Not exactly. The tool extracts the text into an editable document; complex columns, tables and precise fonts may not be reproduced. Plain paragraphs convert most faithfully.

Handwriting is unreliable. The tool is built for printed or on-screen text. Neat capitals may work, but expect more corrections.

Good with clear, well-lit, printed images, but never assume perfection. Lookalike characters and faint print can cause typos, so proofread the result against the image.

Yes. The output is a standard Word document that opens in Microsoft Word, Google Docs, LibreOffice Writer and other compatible apps.

Related Tools