Image to Text (OCR)

Image to Text reads the words in a picture and gives them back as ordinary text. This is called OCR, short for optical character recognition: the computer looks at the shapes of the letters and works out what they say. Drop in a screenshot, a photo of a recipe or a scanned letter, and in seconds you can copy the words instead of typing them.

It is handy for students quoting a textbook, office workers re-using a scanned form, or anyone sent a picture of a document instead of the document.

The reading is done by Tesseract, a well-known open-source OCR engine, inside your browser. Your pictures are never uploaded: only the engine and the language you pick are downloaded, from this site. Each result shows how sure the reader is and boxes every word it found, with unsure words in orange so you know what to check.

1. Tap “Choose files”, drag pictures onto the box, or paste a screenshot with Ctrl+V (⌘V on a Mac). You can add up to 20 at once.
2. Pick the “Language of the text”. For languages other than English, leave “Also read English words” ticked if the text mixes in English.
3. Choose “Keep lines” to keep the line breaks as they are in the picture, or “Join lines” to rejoin each paragraph for pasting into a document.
4. Press “Read text”. Each picture shows its own progress bar, then its text in a box beside it.
5. Check the words boxed in orange (click a box to jump to that word), fix anything that is wrong, then press “Copy” or “.txt”. “Download text” saves everything in one file.

• 19 languages: English, Spanish, French, German, Portuguese, Italian, Dutch, Polish, Turkish, Indonesian, Vietnamese, Russian, Ukrainian, Hindi, Arabic, Chinese (Simplified and Traditional), Japanese and Korean. Only the one you choose is downloaded, between 0.7 and 3 MB.
• Reads JPG, PNG, HEIC (iPhone photos), WebP, AVIF, TIFF, GIF, BMP, SVG and ICO. Camera RAW files and PDFs are not opened directly.
• Up to 20 pictures in one go, read one after another, each with its own progress bar.
• Clean-up for photos of pages: evens out shadows and uneven light, and flips light-on-dark (dark mode) screenshots to dark-on-light before reading.
• Straightens text that is slightly tilted, and a “Turn” button for pictures that are sideways or upside down.
• Small pictures, like most screenshots, are enlarged up to 2× before reading, so small letters read better.
• A confidence badge on every result, boxes around each word found, and unsure words marked in orange.
• Editable results with Copy and Download .txt for each picture, plus Copy all text and one combined download.

• Choose the right language. Reading French text with the English reader loses the accents, and Chinese or Arabic text with the wrong language gives nonsense.
• For photos, hold the phone flat above the page, fill the frame with the text and avoid your own shadow. Sharp focus matters more than anything else.
• Handwriting reads poorly. Neat block capitals sometimes work; joined-up writing usually doesn’t.
• Tables come out as plain lines, not columns. Decorative or tiny fonts, busy backgrounds and curved text are also hard.
• If a screenshot reads worse than expected, open “More options” and try with “Clean up photos” switched off, then read again.