JPG to Word
Extract text from a JPG (JPEG) image with OCR and download it as an editable Word (.docx) document — multiple images become one file. Free and entirely in your browser.
Input
Multiple images become one .docx, each starting on its own page.
Word document
Add one or more JPG images and press Convert to Word. Everything runs in your browser — your images are never uploaded.
Guides
Extract the text from a scanned document, photo or screenshot and download it as an editable Word (.docx) file — right in your browser. No account, no upload to a server, and no watermark.
How to use it
- Drop or choose one or more JPG (JPEG) images (up to 10).
- Pick the image's language.
- Press Convert to Word, then Download the result.
Need OCR on a PNG, GIF, WebP or BMP instead? Use Image to Text Converter — same OCR engine, any image format, plain text output rather than a .docx.
Multiple images become one .docx, each image's text starting on its own page in the order you added them.
How it works
The tool reads the text out of each image with OCR (optical character recognition, via Tesseract.js), then builds a real .docx file from the extracted text — plain paragraphs, no attempt to recreate the original layout, fonts or images.
If you just want the raw text (to paste elsewhere, no .docx needed), use Image to Text Converter instead — this tool exists specifically for the "I need an editable Word file" case.
Private by default
OCR runs entirely in your browser via Tesseract.js, and the .docx is built client-side too — your images are never uploaded anywhere. Only the OCR engine's own files (the recognition model) are fetched from a CDN; your images themselves stay local.
Why isn't the formatting preserved?
OCR reads characters, not layout — it has no concept of the original fonts, columns or images, only the text itself. The output is plain paragraphs you can reformat in Word afterward, not a pixel-for-pixel recreation of the source.
Does it work on handwriting?
No — like most OCR engines, this reads printed/typed text reliably; handwriting recognition is a different, much less accurate problem and isn't supported.
My scan came out garbled — why?
OCR accuracy depends heavily on image quality: a straight-on, well-lit, reasonably high-resolution photo reads far better than a blurry or heavily skewed one. Picking the correct language also matters — OCR only recognizes characters from the language you select.