Convert an Image to Word: Editable Text from a JPG, Photo or Scan
How to convert a picture to text in Word: what Word can do on its own, where it stops, and how to get text from a JPG or photo into Word without uploading it.
Word on a PC can’t turn a picture into editable text by itself. When you add a JPG, a PNG or a photo with Insert › Pictures, that’s exactly what ends up in your document: a picture. To edit the text in it, you need text recognition (OCR) that turns the pixels back into letters. Word only offers a detour through PDF files, and online converters ask you to upload the image.
This guide shows what Word and Windows can do on their own and where that ends, how to get the text from an image into Word directly on your PC, and what to do when the result has errors.
Why a picture in Word isn’t text
A photo, a screenshot or a scanned page is a pattern of pixels to your computer. You see letters, Word sees colored dots. Selecting, searching and correcting only work once text recognition has read the image and output the text as characters.
So “converting an image to Word” usually means one of two things:
- Getting the text out so you can keep working on it in Word. That’s what this guide is about.
- Placing the picture in the document as it is. Word does that on its own: on the Insert tab, Pictures › This device, pick the file, Insert. The text in it stays uneditable.
What Word, OneNote and Windows can do on their own
Microsoft offers three routes, each with limits worth knowing up front.
Word opens PDFs and imports their text. Microsoft’s help page on inserting scanned text describes it like this: scan the document as a PDF with your scanner’s own software, then open the PDF in Word with File › Open; after you confirm with OK, Word imports the text and tries to preserve the formatting. Microsoft names the limits itself:
- This applies to the desktop versions of Word on Windows (Word 2016 through Word for Microsoft 365).
- It’s about PDF files. Word doesn’t open a JPG or PNG as text this way.
- According to Microsoft, the conversion works best with files that are mostly text. If a page is mostly charts or graphics, it may show up as a picture, and the text can’t be edited.
- Accuracy depends on the quality of the scan, and handwriting is seldom recognized, Microsoft says. Proofread the result.
OneNote copies text from pictures. In the desktop app (OneNote for Microsoft 365, 2024, 2021, 2016), right-click a picture you’ve inserted and choose Copy Text from Picture, then paste the text into Word with Ctrl+V. The picture has to be on a OneNote page first, and Microsoft notes that the command isn’t always available right away.
Snipping Tool in Windows 11 recognizes text in a capture. After Win+Shift+S, the Text actions button turns recognition on, and Copy all text puts the text on your clipboard. Microsoft says recognition runs locally on your device. It describes text actions for Windows 11 only and says nothing about whether or how a poor original is prepared first.
Online converters (“JPG to Word”) are the fourth route. They require you to upload the image. For a scanned contract, an invoice with bank details or a doctor’s letter, that’s not a small thing.
Convert an image to Word with Celerando: paste the text, no upload
Text recognition in Celerando reads whatever is visible on your screen. You open the image in any program, drag a frame over it, and the recognized text is on your clipboard. In Word, you paste it with Ctrl+V. It doesn’t matter whether the image is a JPG, a PNG, a photo from your phone, a screenshot or a page in a PDF.
Recognition runs on your computer, with Tesseract 5. Neither the image nor the recognized text leaves it. Before recognition, Celerando prepares the original in six stages (scaling, greyscale, contrast, binarisation, sharpening, noise reduction), bundled in three profiles. It recognizes 14 languages, several at once. Celerando runs on Windows 10 and 11.
Edition limits: text recognition is included from Celerando Pro (€36 one-time, no subscription). You can try all features free for 14 days; afterwards, Celerando keeps running as Free, without text recognition. All editions are listed under Editions and prices.
Step by step
- Install Celerando. The installer is signed and needs no admin rights and no account.
- Open the image and show it large enough that the text is easy to read, for example in the Photos app, in your browser or in a PDF viewer. The bigger the letters on screen, the better the result.
- Pick the right profile under Configuration → OCR → OCR profile: Photo / camera for photographed pages with uneven light, Scan / document for scans and paper, Screenshot for screenshots. You set this once, not for every image.
- Press Ctrl+Alt+T and drag a frame over the text. The frame can span several monitors. You can change the hotkey in the configuration.
- Press Ctrl+V in Word. The text is in your document and editable, with no saving in between and no export.
For several images or pages, repeat steps 4 and 5 for each page. If the image mixes two languages, such as German and English, Celerando recognizes it in one pass; that’s the default language selection.
What doesn’t come along
Celerando puts the recognized text on your clipboard, not the layout. You rebuild headings, bold text, columns and table borders in Word. With tables, it helps to capture the columns one at a time instead of dragging one frame over the whole table. Logos, photos and graphics in the image stay pictures; add them separately with Insert › Pictures if you need them.
If the exact look matters, say for a form that should look just like the original, the PDF route in Word is worth a try: Word tries to preserve the formatting there, with the limits Microsoft itself names above.
When it doesn’t work
- The recognized text has errors. Zoom in until the letters are crisp and drag the frame again. Tiny type is the most common cause of garbled letters. Check the profile: a photo with shadows needs Photo / camera, not Scan / document. In the manual Pro profile, you control the six stages yourself and can enlarge the original up to 3× before recognition.
- Accents or special letters are missing. The language is usually missing from the selection. Eight languages come bundled; six more (Danish, Swedish, Finnish, Norwegian, Greek, Russian) you download under Configuration → OCR → Text recognition languages, one model of a few megabytes each.
- Ctrl+Alt+T does nothing. After the 14 trial days, Celerando runs as Free, and text recognition is included from Pro. If another program uses the same shortcut, set a different one in the configuration.
- Word opens the PDF, but the text is a picture. Microsoft describes this for pages that are mostly graphics. Open the PDF in a PDF viewer instead and get the text out with text recognition. Why text in some PDFs can’t be copied at all is covered in the guide Can’t copy text from a PDF?.
- Text actions is missing in Snipping Tool. Microsoft describes text actions for Windows 11 only. On Windows 10, use OneNote or Celerando.
- Handwriting comes out as nonsense. Microsoft says handwriting is seldom recognized. Celerando also describes its text recognition for scans, photos and screenshots, not handwriting; test it with your own original during the trial before you rely on it.
Celerando and Word compared
| Celerando | Word (PDF via “File › Open”) | |
|---|---|---|
| Source material | Anything visible on screen: JPG, PNG, photo, screenshot, PDF page | PDF files; a JPG or PNG stays a picture in Word |
| Result | Plain text on the clipboard, you rebuild the layout | Word tries to preserve the formatting |
| Pages with lots of graphics | Drag the frame over the text only | Page may show up as a picture, text not editable |
| Image preparation | Six stages, three profiles, upscaling 1.5× in Screenshot and Scan / document (up to 3× in the manual Pro profile) | Not described by Microsoft |
| Languages | 14, several at once | Not stated on Microsoft’s help page |
| Windows and programs | Windows 10 and 11, any target program | Desktop versions of Word on Windows, Word 2016 and later |
| Price | Part of Pro, €36 one-time, 14-day trial | Included in Word |
As of November 2026. OneNote (“Copy Text from Picture”) and Snipping Tool (“Text actions”, Windows 11 only) also recognize text without uploading a file; Microsoft describes no preparation of poor originals for either, and in OneNote the picture has to be on a notes page first. Online converters require you to upload the image.
Frequently asked questions
Can Word convert a picture to text?
Not directly on a PC. A JPG or PNG added with “Insert › Pictures” stays a picture. Word only imports the text of PDF files you open with “File › Open”, and according to Microsoft that works best with files that are mostly text. For pictures you need text recognition, such as Celerando, OneNote or Snipping Tool in Windows 11, and then paste the text into Word.
How do I convert a JPG to Word for free?
For free, you can use OneNote (“Copy Text from Picture”) or, in Windows 11, Snipping Tool (“Text actions”), both without uploading, or an online converter that uploads the image. In Celerando, text recognition is included from Pro (€36 one-time) and free to try for 14 days; the free edition doesn’t include it.
How do I get the text from a phone photo into Word?
Send the photo from your phone to your PC, open it on screen, and drag the text recognition frame over it. The Photo / camera profile is meant for photographed pages with uneven light. Shoot as straight from above as you can, with even light.
Is the formatting preserved?
With Celerando, no: you get the text, and you set headings, bold text and tables again in Word. If the original is a PDF, Word tries to preserve the formatting when it opens it, though Microsoft says the result might not look exactly like the PDF.
Do I have to upload the image anywhere?
No. Celerando recognizes text on your computer; with the eight bundled languages, text recognition needs no connection at all, which you can try in flight mode. Celerando only connects to the internet for the downloads you choose (models, extra languages, the optional CUDA runtime), the update check, license activation including the query of seats in use, a one-time installation ping on first launch, and addresses you enter yourself (web app tiles load the site’s icon, the “Connections” tile checks your servers). Your image is never part of this.
Can I translate the text from the image right away?
Yes. How to recognize text in an image and translate it with Celerando’s local translator is covered in the guide Translate text in an image.
Sources
- Microsoft Support: Insert scanned text or images into Word (inserting pictures, PDF via “File › Open”, accuracy, handwriting; desktop versions of Word on Windows; accessed November 13, 2026)
- Microsoft Support: Opening PDFs in Word (works best with files that are mostly text, pages with graphics as a picture; accessed November 13, 2026)
- Microsoft Support: Copy text from pictures and file printouts using OCR in OneNote (accessed November 13, 2026)
- Microsoft Support: Use Snipping Tool to capture screenshots (Text actions, Copy all text, local text recognition, Windows 11 section; accessed November 13, 2026)
More guides
- Print Screen not working? Fix PrtScn and Snipping ToolPrtScn, Win+Shift+S or Snipping Tool doing nothing? The usual causes on Windows 10 and 11, a way to screenshot without Print Screen, and what actually helps.
- How to Edit a Screenshot on Windows: Crop, Highlight, Add TextCrop a screenshot, highlight something, add an arrow or text: what the built-in tools do on Windows 10 and 11, where they stop, and a faster way.
- How to screenshot one monitor on Windows 11 and 10Two monitors, but the screenshot should show only one? What each Windows screenshot key captures and how to grab exactly one screen.