How to Extract Text from an Image (Every Device, Free)

Text in a picture looks like text, but to your computer it is just coloured pixels. You cannot select it, search it or paste it anywhere. Getting it out takes OCR (optical character recognition): software that looks at the shapes and works out which letters they are.

The good news is that you almost certainly have OCR already, on your phone and probably on your computer, and a free browser tool covers everything else. The bad news is that OCR makes mistakes, and the dangerous ones are not the garbled words you will notice. They are the numbers that come out wrong but still look right.

We made nine test images (a clean screenshot, tiny text, a tilted photo, a noisy blurred scan, low contrast, dark mode, two columns, accented French and German, and a line of code) and ran every one through FileNaut's Image to Text (OCR) tool and through Apple's built-in text recognition. This guide shows the quickest method on each device and what our tests say you should check before you trust the result.

• Any computer, any browser: drop or paste the image into FileNaut OCR, click Extract Text, then Copy or download it as .txt or Word.
• iPhone or iPad: open the photo, tap the Live Text button, then Copy All.
• Android: open the photo in Google Photos or Google Lens and tap Text, then Copy.
• Mac: open the image in Preview or Photos and drag across the text as if it were a document.
• Windows 11: take a screenshot with the Snipping Tool and click Text actions, then Copy all text.

Whatever you use: check every number against the original before you paste it anywhere that matters.

This works the same on Windows, Mac, Linux and Chromebook, and the image never leaves your device: the recognition runs inside your browser rather than on a server.

1. Open the Image to Text tool.
2. Add your image. You can drag it in, click Browse Files, or paste a screenshot straight from your clipboard with Ctrl+V (⌘+V on a Mac). JPG, PNG, WebP, BMP, GIF and iPhone HEIC photos all work.
3. If the text is not in English, choose its language from the menu next to the button. Seventeen languages are available, including Spanish, French, German, Russian, Arabic, Hindi, Chinese, Japanese and Korean.
4. Click Extract Text. The first run downloads the recognition engine and language data, so it takes a few seconds; after that, a screenshot-sized image finishes in about a second.
5. Click Copy, or download the result as a .txt file or a Word document.
6. Read the result against the image before you use it. The next sections explain what to look for.

Have a scanned PDF rather than an image? Use PDF OCR, which reads every page of the document in one go. Our guide to converting a PDF to text explains how to tell whether a PDF needs OCR at all.

Since iOS 15, iPhones recognise text in photos automatically.

1. Open the photo or screenshot in the Photos app.
2. Tap the Live Text button (a small box with lines in it) in the bottom-right corner. The text in the photo is highlighted.
3. Tap Copy All, or press and hold a word and drag the handles to copy just part of it.

You can do the same thing without taking a photo: point the Camera at the text, tap the Live Text button when it appears, and copy. Live Text uses Apple's recognition engine, which did very well in our tests (see the comparison below).

1. Open the image in Google Photos and tap Lens, or open the Google app and tap the camera icon in the search bar.
2. Tap Text. Lens highlights everything it can read.
3. Tap Select all, then Copy text. You can also send it straight to a computer signed in to the same Google account.

On many recent phones you can also long-press a screenshot preview and select text directly. Menu names vary between manufacturers, but Google Lens is available on almost every Android phone.

On macOS Monterey or later, text inside images is selectable.

1. Open the image in Preview (double-click most image files) or in Photos.
2. Move the pointer over the text until it turns into a text cursor.
3. Drag across the text to select it, then press ⌘+C.

If nothing selects, Live Text may be switched off: in Photos, turn on Settings → General → Enable Live Text.

Windows 11 has three built-in or free options.

• Snipping Tool: press Win+Shift+S, capture the area with the text, open the capture in the Snipping Tool and click Text actions, then Copy all text. This needs an up-to-date Windows 11.
• PowerToys Text Extractor: install Microsoft's free PowerToys, press Win+Shift+T and drag a box over any text on screen. It goes straight to your clipboard.
• OneNote: paste the image into a page, right-click it and choose Copy Text from Picture.

For an image file that is already saved, the quickest route on Windows 10 or 11 is usually to drop it into the online OCR tool, which also gives you a .txt or Word download.

1. Upload the image to Google Drive.
2. Right-click it and choose Open with → Google Docs.
3. A new document opens with the image at the top and the recognised text underneath, ready to edit.

This keeps paragraphs reasonably well and is handy when you want an editable document rather than a clipboard. It does upload the image to Google, so skip it for anything private.

We ran the same nine images through the browser tool (which uses the open-source Tesseract engine) and through Apple's built-in recognition, the engine behind Live Text on iPhone and Mac.

Three lessons come out of this:

• Wrong numbers look right. 604.555.0190 and 504-555-0199 are perfectly believable phone numbers. Nothing about them signals an error, which is why every number, date, amount and email address needs checking by eye.
• Screenshots are easy; photos are hard. Both engines were near-perfect on clean digital text, including low-contrast and dark-mode screenshots. Tilt and small text are what break recognition.
• Phone engines are stronger on difficult photos. For a skewed photo of a page, Live Text or Google Lens is the better first try. For screenshots, saved images and anything on a desktop, the browser tool is just as accurate and gives you a file.

• Shoot straight on. Hold the phone parallel to the page. Our 8° tilt was enough to change a digit in a phone number.
• Get closer instead of zooming later. Text should be large in the frame. If letters are only a few pixels tall, no engine can rebuild detail that is not there. Enlarging a small image with an image resizer sometimes helps a little, but a closer photo helps much more.
• Crop to what you need. One column at a time avoids the left-right mix-up, and cutting away logos and photos removes stray characters.
• Pick the right language. With the wrong language selected, accents vanish and letters like ß turn into nonsense.
• Even light, no shadows. A shadow across a line looks like a smudge to OCR.
• Expect handwriting to be poor. Engines like Tesseract are built for printed text. Neat block capitals sometimes work; joined-up handwriting usually does not. Google Lens copes noticeably better with handwriting than browser tools.

It depends where the recognition happens. Many "image to text" websites upload your image to their server, process it there and send the text back, which means a copy of your receipt, ID card or private message sits on someone else's computer.

FileNaut's OCR tool does the recognition inside your own browser. The engine and language files are downloaded to your device the first time you use it, and your image is never sent anywhere. Phone methods (Live Text on iPhone, and on-device recognition on newer Android phones) also work locally. Google Docs and the cloud features of Google Lens do send the image to Google.

New to the technology? What is OCR? explains how it works.

On a phone, use what is built in: Live Text on iPhone, Google Lens on Android. On a computer, paste or drop the image into the Image to Text tool, pick the language and copy the result. Shoot straight on and close, crop to one column, and before you use the text anywhere it matters, check every number against the picture. OCR's worst mistakes are the ones that look correct.