Image to Text (OCR)
Free online Image to Text (OCR). No sign-up needed, and it runs in your browser.
Copy the text out of a photo, screenshot or scanned page with OCR, in 20+ languages. Paste a screenshot or drop up to 20 pictures.
Tip: you can also paste a screenshot here (Ctrl+V / ⌘V).
About Image to Text (OCR)
Retyping text from a photo of a sign, a screenshot of a chat or a scanned letter is slow and easy to get wrong. OCR reads the words for you in 22 languages, including Arabic, Chinese, Hindi, Russian and Urdu, and handles up to 20 pictures in one go. You can copy the result, save it as a .txt file or make a searchable PDF, and small pictures are enlarged automatically to help recognition.
How to use Image to Text (OCR)
- Click Browse or drop up to 20 photos, screenshots or scans, or paste a screenshot with Ctrl+V.
- Pick the Language of the text, and keep Also English (for mixed text) ticked if English words appear too.
- Tick Also make a searchable PDF if you want one, then click Read text.
- Check the words in Text found, then click Copy text or Download text (.txt).
How your data is handled
Your pictures are read in your browser and never uploaded. The first time, the text-recognition engine (about 4 MB) and language data (2–5 MB per language) download from jsDelivr, a public code CDN, and your browser keeps them.
Limits and accuracy
- Handwriting, decorative fonts and text on busy backgrounds are often misread, so always check names and numbers.
- The result is plain text; tables, columns and formatting aren't kept.
- Turn sideways or upside-down pictures upright first with the Photo Rotator.
- The first run needs an internet connection to download the engine and language data.
Questions and answers
Which languages can it read?
It reads 22 languages, including English, Arabic, Chinese (Simplified and Traditional), French, German, Hindi, Japanese, Korean, Persian, Russian, Spanish, Urdu and Vietnamese. For text that mixes in English words, keep Also English (for mixed text) ticked.
Are my pictures uploaded to a server?
No. The text is read in your browser, on your own device. Only the reading engine and language data are downloaded the first time, from jsDelivr, and your browser keeps them so later visits start faster.
Can it read handwriting?
It's built for printed and on-screen text. Very neat handwriting may partly work, but expect mistakes, so check every line or type it out yourself.
What is the searchable PDF?
Each picture becomes a PDF page with the recognised words behind it as an invisible text layer. The page looks like your picture, but you can search, select and copy its text in any PDF reader.
Why is some of the text wrong?
Blur, small print, tilt, poor light or the wrong language setting all cause mistakes. Crop to the text, use a sharp, straight picture, and check the Average confidence: under 60% usually means the picture needs improving.