# Convert Images to Text in Word: What Works, What Fails

- Author: Schema Law (https://www.schemalaw.com/)
- Published: 2026-05-24T12:06:52Z
- Updated: 2026-10-11T19:22:20Z
- Category: OCR in everyday software
- Canonical: https://www.schemalaw.com/blog/convert-images-to-text-in-word

> Convert images to text in Word by saving the picture as a PDF and reopening it, or by copying the text out with OneNote or a screenshot tool. Steps inside.

Word has no documented command that will convert images to text on the page. A picture you insert stays a picture. To get editable words, either save the picture as a PDF and open that PDF in Word, or copy the text out of the image with OneNote or a screenshot tool and paste it into Word.

Which of those is worth your time depends on the image. A screenshot of an email, a phone photo of an exhibit and a scanned letter saved as a JPG are three different problems. The sections below take the routes one at a time and end with a table for choosing.

## Can Word convert an image to text by itself?

Not from a picture sitting in a document. Microsoft's Word help describes two separate tasks: inserting a scanned image, which leaves you with a picture, and scanning text, which goes through a PDF that Word converts when it opens the file. It documents no button that turns an inserted picture into words.

You can see the division on Microsoft's page about [inserting scanned text or images into Word](https://support.microsoft.com/en-us/office/insert-scanned-text-or-images-into-word-b4ae150e-319f-4e18-b27b-418f1d690823). Under "Scan an image into Word" the steps end with Insert, Pictures. Under "Scan text into Word" the steps go a different way: get a PDF, then use File, Open so that Word imports the text.

So the honest short answer has two parts.

- **Directly: no.** Microsoft's pages on pictures in Word cover inserting, resizing and wrapping them, not reading them.
- **Indirectly: sometimes.** Word's text recognition runs when Word opens a PDF for conversion. If you can turn your image into a PDF, you can put it in front of that conversion.

The other routes in this post do the reading outside Word and use Word only as the place where the text ends up.

Before you start on any of them, ask whether the image is the only copy. If the screenshot shows an email, the email itself has real text. If the JPG is a scan of a letter someone still holds, a fresh scan to PDF is a better source than any picture. Converting an image is the fallback for when the original text is out of reach.

## What happens to a JPG or PNG pasted into a Word document?

It stays a picture. Word places the image on the page as an object you can resize, move and wrap text around. The words in it are pixels. You cannot click into them, search for them, spell-check them or copy them, however sharp the image looks.

Microsoft's page on [inserting pictures](https://support.microsoft.com/en-us/office/insert-pictures-3c51edf4-22e1-460a-b372-9329a8724344) describes what you can do with one: select Insert, Pictures, This Device to add it, drag a corner handle to resize it, and pick a wrapping option to control how body text flows around it. Every one of those is about the picture as a shape.

### How to tell what you have

Sometimes a document reaches you and you cannot tell whether a block of words is text or a picture of text. Three quick tests:

1. **Click once on a word.** If a frame with handles appears around the whole block, it is a picture.
2. **Press Ctrl+F and search for a word you can see in the block.** No match means there is no text there.
3. **Try to change the font.** Select the area and pick another typeface. A picture ignores you.

### Why this matters in legal papers

A picture of text inside a Word document causes problems that only show up later.

- **Search misses it.** A colleague searching the matter folder for a party's name will not find the declaration where that name appears only inside a pasted screenshot.
- **Word count skips it.** If a court limits your brief by word count, words inside an image are not counted by Word. Whether the court counts them is a question for that court's rules.
- **Screen readers cannot read it** unless you add a description.
- **It will not reflow.** Change the margins and the picture's line breaks stay where they were.

None of that means a picture is the wrong choice. When the point is to show what a document looked like, such as a text message exchange or a signed page, the picture is the evidence and should stay a picture. Add the words as real text nearby if you also need them to be searchable.

When the words come from an exhibit photo, check the paste for the invoice number, a name taken from the affidavit, and a date on the pleading. If the picture is a medical record, find the clinic name and the visit date in the text and on the photo. Fix every mismatch before the draft leaves your desk.

## Saving the picture as a PDF and reopening it in Word

The Word-only route has three steps: put the picture alone on a page in a blank document, save that document as a PDF, then open the PDF from inside Word and accept the conversion. Microsoft documents each step separately. It does not document the combination, so test it on your image before you count on it.

### The steps

1. Open a blank Word document.
2. Go to **Insert**, **Pictures**, **This Device** and insert the image. Drag a corner so it fills the page width.
3. Save the document as a PDF. Microsoft's [Edit a PDF](https://support.microsoft.com/en-us/office/edit-a-pdf-b2d1d729-6b79-499a-bcdb-233379c2f63a) page gives the path as **Export**, **Create PDF/XPS Document**, **Create PDF/XPS**, then **Publish**.
4. Close the document.
5. Go to **File**, **Open**, and open the PDF you just made.
6. Word says it will make a copy of the PDF and convert its contents. Select **OK**.

### What to expect

There are two possible outcomes, and Microsoft's documentation describes both.

The first is text. Microsoft's scanning page says that when Word opens a scanned PDF it imports the text, and that the accuracy of the text recognition depends on the quality of the scan and the clarity of the text.

The second is your picture, back again. Microsoft's page on [opening PDFs in Word](https://support.microsoft.com/en-us/office/opening-pdfs-in-word-1d1d2acc-afa0-46ef-891d-b76bcd83d9c8) says that if the PDF contains mostly charts or other graphics, the whole page might show up as an image, and then the text can't be edited. A PDF made from one large picture is, by construction, mostly graphics. Microsoft's pages talk about PDFs that came from a scanner and do not say how Word treats a PDF you built from a JPG, so the only way to know is to try.

### Improving the odds

These are general points about text recognition, not promises about Word.

- Use the original image file, not a copy that has been through a messaging app, which often shrinks pictures.
- Crop away everything that is not the text: desk, fingers, browser toolbars.
- Keep the page upright. Rotate the picture in Word before saving if it came in sideways.
- One image per page.

If the first attempt returns a picture, do not spend an afternoon adjusting. Move to one of the other routes.

### If the image is still on paper somewhere

When the "image" is really a scan of a paper page, and the paper is available, skip the picture entirely. Scan the page to a PDF and open that in Word, which is the method Microsoft actually recommends for text.

## Routing a picture through OneNote on the way to Word

OneNote can pull the text out of a picture and put it on the clipboard, and you paste it into Word. You insert the image on a OneNote page, right-click it, choose Copy Text from Picture, switch to Word and paste. Microsoft says the pasted text can go into another program such as Word.

The post on OneNote covers the feature in full, including printouts, the Mac version and the waiting period, in its section on [Copy Text from Picture, OneNote's OCR for scanned pages](https://www.schemalaw.com/blog/microsoft-onenote-handwriting-to-text#copy-text-from-picture-onenotes-ocr-for-scanned-pages). What follows is only the trip from image to Word.

### The trip

1. Copy the image, or have the file ready.
2. In OneNote, click on a blank page and paste, or insert the file.
3. Right-click the image and choose **Copy Text from Picture**.
4. In Word, click where the words belong and press Ctrl+V.
5. From the paste options choose **Keep Text Only**, which Microsoft's page on [paste formatting in Word](https://support.microsoft.com/en-us/office/control-the-formatting-when-you-paste-text-20156a41-520e-48a6-8680-fb9ce15bf3d6) says discards formatting and gives the text the style of the paragraph it lands in.

### Two things that catch people

**The command may not be there yet.** Microsoft's page on [copying text from pictures in OneNote](https://support.microsoft.com/en-us/office/copy-text-from-pictures-and-file-printouts-using-ocr-in-onenote-93a70a2f-ebcd-42dc-9f0b-19b09fd775b4) says the command may not be available immediately, and the OneNote post gives the waiting time Microsoft states. If you need the words in the next ten minutes and the menu item is missing, use a screenshot tool.

**You receive words, not layout.** Microsoft describes copying text. Expect to rebuild paragraphs, lists and any table yourself once the text is in Word, and to read it against the image.

### Tidy-up in Word

After pasting, turn on formatting marks, join lines that were split mid-sentence, and reapply your document's styles. Then put the image and the pasted text side by side and check names, numbers and dates first.

## Screenshots of text: the Windows and Mac tools that copy it

Both operating systems can copy text out of a screenshot without Word or OneNote. On Windows 11, Snipping Tool has a Text actions button that recognizes the text in a snip. On a Mac, Live Text in Preview lets you select and copy text in an image. Paste the result into Word.

### Windows: Snipping Tool

Microsoft's page on [using Snipping Tool to capture screenshots](https://support.microsoft.com/en-us/windows/apps/use-snipping-tool-to-capture-screenshots) gives the shortcut Windows logo key + Shift + S to open the snipping overlay. Once you have captured a snip, you select the **Text actions** button, which the page says activates optical character recognition. From there you can select and copy specific text or use **Copy all text**.

Two details on that page are useful in a law office.

- Microsoft says all text recognition processes are performed locally on your device.
- A **Quick redact** tool hides email addresses and phone numbers in the snip. That affects the screenshot image. It is not a substitute for [proper redaction](https://www.schemalaw.com/blog/ocr-before-or-after-redaction) of a document you will produce.

The page has separate instructions for Windows 11 and Windows 10. Text actions is described in the Windows 11 instructions, and the Windows 10 instructions do not mention it. If you do not see the button, check which version of Windows and Snipping Tool you are running.

### Mac: Live Text in Preview

Apple's Preview guide explains how to [interact with text in a photo using Live Text](https://support.apple.com/guide/preview/interact-with-text-in-a-photo-prvw625a5b2c/mac). You open the image in Preview, hold the pointer over the text, drag to select it, Control-click the selection and choose **Copy Text**. Apple notes that if you can't select text, you should choose Tools, Text Selection, and that Live Text isn't available in all regions or languages.

Live Text copies words out of an image. It is a separate question whether a Mac can write those words back into a PDF as a searchable layer, and the post on [converting a PDF to a searchable PDF on a Mac](https://www.schemalaw.com/blog/convert-pdf-to-searchable-pdf-mac) takes that up.

### A screenshot is often the wrong starting point

If you are taking a screenshot of something on your own screen in order to convert it, stop and look for the text itself. A web page, an email and most PDFs let you select and copy directly. Screenshot tools earn their place when the source will not let you select: a video frame, a scanned PDF in a viewer, a remote desktop session, an image someone else sent.

## Which route suits an exhibit photo, a screenshot or a scanned letter?

Match the route to the image. A screenshot goes through the operating system's own tool. A scanned letter is better rescanned to PDF and opened in Word or run through a converter. A phone photo of an exhibit needs the most care, because the photo itself is usually the weak point.

| The image | First choice | Why | Fallback |
| --- | --- | --- | --- |
| Screenshot of an email, web page or chat | Snipping Tool Text actions, or Live Text in Preview | Short text, sharp pixels, nothing to install | OneNote's Copy Text from Picture |
| Scanned letter saved as a JPG or PNG | Rescan the paper to PDF and open it in Word | It is the route Microsoft documents for text | Save the image as a PDF and reopen it in Word, or a converter that accepts images |
| Phone photo of a paper exhibit | Retake or scan it properly, then convert | Angle, shadow and blur all work against recognition | A converter that accepts JPEG, then proofread closely |
| Photo of a table, such as a statement or schedule | A converter that outputs tables | Copy-text tools are documented as copying text, with no promise of table structure | Type the table from the image |
| Picture of handwriting | Type it | Microsoft says handwritten text is seldom recognized | Test one page in the tool you have |
| Image that is itself the evidence | Leave it as a picture and add a typed transcription | The look of the original matters | None needed |

For phone photos, the guide on [how to OCR a photo of a document](https://www.schemalaw.com/blog/how-to-ocr-a-photo-of-a-document) covers how to take a picture that software can read and what to do with one that is already taken.

### A worked example

Tomas Leclerc is preparing a declaration in a dispute between Garnet Supply and a former distributor. He has three images:

1. A screenshot of a pricing email the client forwarded as a picture. He opens it, uses Text actions in Snipping Tool, copies the two paragraphs he needs and pastes them into Word.
2. A phone photo of a one-page delivery receipt with a small table. The copy-text tools are documented as copying text, not table structure, so he asks the client to bring the paper in, scans it to PDF and converts that.
3. A photo of the distributor's handwritten note. He leaves it as a picture, attaches it as an exhibit and types a transcription beneath, marked as his own reading.

In each case he compares the typed words with the image before the draft leaves his desk.

### When the pictures keep coming

If images of documents arrive every week, a repeatable process beats a clever workaround. Schema Law accepts PNG and JPEG files as well as PDFs, up to 20 MB each, reads them into text and tables, and exports Word. Before you send any client image to an online service, this one included, read what it says about your files. Schema Law's commitments are in the [security section of the homepage](https://www.schemalaw.com/#security). Then test it on your own images and check the output, the same as you would with the Word, OneNote and screenshot routes.

## Try it on a real document

Upload a scanned PDF and export it as Word, a searchable PDF or plain text on the free plan. Your documents are never used to train an AI model. [Start free](https://www.schemalaw.com/api/auth/google)