How to Copy Text From a Screenshot or Photo

To copy text from a screenshot or photo, run it through OCR (optical character recognition), which reads the letters in an image and turns them into text you can edit. With ImageSpark's Image to Text, you paste the screenshot with Ctrl+V (Command+V on a Mac), choose the language, click Extract Text and copy the result. The OCR engine runs inside your browser, so the image is never uploaded.

This guide covers that method step by step, the built-in options on phones and computers, and how to get accurate results from difficult images.

How OCR works, in brief

An OCR engine looks for shapes that resemble letters, groups them into words and lines, and matches them against what it knows about a language. Clear, high-contrast, printed text is easy. Blurry, tilted, handwritten or decorative text is much harder. That's why the quality of the image matters more than anything else, and why a few seconds spent cropping and straightening pays off.

Copy text from an image in your browser

1. Open the tool. Go to Image to Text.

2. Add the image. Paste a screenshot straight from your clipboard with Ctrl+V or Command+V, or use the Paste button. You can also drop a file or choose one. JPG, PNG, WebP, AVIF, GIF and BMP images all work.

3. Pick the language. Choose the language of the text. Twelve are available: English, Spanish, French, German, Italian, Portuguese, Dutch, Chinese (Simplified), Japanese, Arabic, Hindi and Russian. If the text mixes one of these with English, turn on the option to also recognize English.

4. Extract the text. Click Extract Text. The first time, your browser downloads the OCR engine and the language data, so it takes a little longer. Your browser keeps them afterward, so later runs start faster.

5. Check, edit and copy. The text appears in an editable box. Fix any mistakes, then copy it to the clipboard or download it as a .txt file.

Built-in options on phones and computers

Many devices can already pick out text from what's on screen:

iPhone, iPad and Mac: Live Text lets you press and hold, or click and drag, on text in photos and screenshots to select and copy it.
Android: many phones offer a text-selection option in the camera or gallery app when you point it at text.
Windows 11: the Snipping Tool can copy text from a capture using its text actions.

These are great for a quick copy of what's on your screen. A browser tool is handy when you're working with image files rather than the screen, when you want to edit and save the text as a file, or when you need one of several languages. With Pro, you can also queue a batch of images and get all their text in one file, with a heading for each image.

Tips for accurate results

Use the original screenshot. A screenshot is pixel-perfect. A phone photo of a computer screen adds glare, moiré patterns and blur, so avoid it if you can.
Crop to the text you need. Cropping away toolbars, pictures and other clutter gives the engine less to misread. Use Crop Image and keep just the relevant block.
Turn the image upright. OCR works best on level text. Straighten a sideways or tilted image with Rotate & Flip before extracting.
Make small text bigger before capturing. If you're taking a screenshot of tiny text, zoom in on the page first so the letters are larger.
Photograph pages in good light. Hold the phone parallel to the page, avoid shadows from your hand, and make sure the photo is sharp.
Choose the right language. Accents and characters are read far more accurately when the engine knows which language to expect.
Proofread numbers and codes. OCR can confuse similar characters such as O and 0, or l and 1. Always double-check account numbers, addresses and passwords.

What OCR struggles with

Even a good engine has limits. Expect to do some correcting with:

Handwriting. Neat block capitals may work, but joined-up writing usually produces errors.
Decorative or unusual fonts, such as script lettering on a poster.
Text over busy backgrounds, like a caption on a photo of foliage.
Tiny or blurry text, including heavily compressed images saved from chat apps.
Curved pages, such as the area near the spine of a thick book.
Tables and columns. The words come out, but the layout is flattened into lines of text, so you may need to rebuild the table.

Photos of documents and receipts

For a printed page, a receipt or a business card, the best results come from a flat, well-lit photo taken from directly above. Fill the frame with the document and crop away the table around it. For long receipts, photograph them in sections and extract each part separately rather than shrinking the whole thing into one image where the text becomes tiny.

If your photos are iPhone HEIC files, convert them first with HEIC to JPG, since the OCR tool reads the common web formats.

Privacy: screenshots often contain personal data

Think about what's in the screenshots you want to read: chat messages, bank balances, email threads, invoices, customer details. Uploading them to an unknown server to extract a few lines of text is a real privacy trade-off. ImageSpark's OCR runs on your own device, so the image stays with you. Only the engine and language data are downloaded to your browser.

FAQ

Can OCR read handwriting?

Only to a limited extent. OCR engines are trained mainly on printed and on-screen text, so neat block capitals may come out fine while cursive usually needs heavy correcting.

Which languages can Image to Text read?

English, Spanish, French, German, Italian, Portuguese, Dutch, Chinese (Simplified), Japanese, Arabic, Hindi and Russian, plus an option to recognize English alongside any of them.

Why is some of the extracted text wrong?

The image may be blurry, tilted, low contrast or set in an unusual font, or the wrong language may be selected. Crop, straighten and use the sharpest version you have, then try again.

Is my image uploaded to extract the text?

No. Recognition runs in your browser on your own device. Only the OCR engine and language data are downloaded, and they're cached for next time.