Image to Text (OCR). Words out of pixels, privately.

How to use the Image to Text (OCR) tool

This tool reads the words in a photo, scan or screenshot and gives them back as text you can edit, copy or save. It runs the open source Tesseract.js recognition engine inside your browser, so the image is read on your own device. It is handy for copying text out of screenshots, receipts, printed pages and slides.

Step by step

1. 1
   Load an image

   Drop a JPG, PNG, WebP or GIF on the card that says Drop an image or screenshot, or click the card to browse. On a keyboard, focus the card and press Enter. You can also paste a screenshot anywhere on the page with Ctrl + V, or Cmd + V on a Mac. One image is read at a time.

2. 2
   Pick the language

   Choose the language of the text from the Language menu: English (the default), Spanish, French or German. Picking the right one helps the engine with accents and common words.

3. 3
   Extract the text

   Press Extract text. The button reads Working… and a progress bar shows each stage. The first run downloads the engine and the language data, which the page puts at about 10 to 15 MB, so it takes longer than later runs.

4. 4
   Check and correct the result

   The text appears in the Extracted text box next to your image. The box is editable, so fix any misread words there. The word and character count under it updates as you edit.

5. 5
   Copy or save it

   Press Copy text to put the text on your clipboard, or Download .txt to save it as a text file named after your image. Press Clear to remove the image and the text and start again.

Reading the result

• The confidence figure next to Extracted text, shown as confidence ~ and a percent, is the engine's own estimate of how sure it is. A low figure means you should proofread closely.
• Line breaks follow the lines in the image, so text from narrow columns comes out with short lines. Join them in the box before you copy if you need flowing paragraphs.
• If the message No readable text found in this image appears, the engine saw no text it could read. Try a sharper or larger image, or crop to just the text.

Tips and common mistakes

• Sharp, straight, high contrast text reads best. Crop out photos, logos and backgrounds before loading the image.
• Very small text reads poorly. Zoom in before you take a screenshot so the letters are larger in the image.
• Handwriting, decorative fonts and text set at an angle often come out wrong. Expect to fix those by hand.
• If the engine fails to load, the page says to check your connection. The first run needs internet access to fetch the engine; later runs can work offline from the browser cache.

Privacy and limits

Your image is read on your device and is not uploaded. The first time you press Extract text, the browser downloads the Tesseract.js engine and the language data from the jsDelivr CDN (cdn.jsdelivr.net) and keeps the language data in its own storage for next time. Nothing else is sent, and the image and text are not saved.

Back to the tool

Everything the tool reads for you.

• Extract text from any photo or scan without uploading it, the OCR runs on your own device.
• Paste a screenshot straight in with Ctrl / ⌘ + V and turn it into editable text instantly.
• Choose your language, English, Spanish, French or German recognition built in.
• Edit the extracted text in the output box, with live word and character counts, before you save it.
• Copy to clipboard or download a .txt file named after your original image.
• Work offline after the one-time engine download, with no data ever leaving your device.

Frequently asked questions.

More handy tools.

Browse the full free-tool library →