Image to Text (OCR)
Extract text from photos and screenshots — recognized entirely in your browser
1. Choose an image
2. Language
Language data downloads once on first use (your image never leaves the device).
3. Extract
Extract text from photos, screenshots, and scanned documents. Supports Korean+English, Japanese, and Chinese — recognition runs entirely in your browser (tesseract), so images never leave your device.
What you can do
- ✓Korean+English·Japanese·Chinese
- ✓Copy · save as TXT
- ✓on-device recognition
- ✓Language data downloads once
How to use
- 1Upload an image (paste works too)
- 2Pick the language
- 3Press Extract, then copy or save
Supported formats · JPG · PNG · WebP · BMP → TXT
Pull the text out of photos, screenshots, and scanned images to copy or save as TXT. Choose Korean+English, English, Japanese, or Simplified Chinese; recognition runs inside your browser, so images are never uploaded. Language data is downloaded only once.
How to use it
- 1Drop or select an image containing text.
- 2Choose the language and click Recognize. The first run downloads language data, so allow a moment.
- 3Review the extracted text, then copy it or save as TXT.
When it comes in handy
Moving a name and phone number from a business-card photo into your contacts, grabbing the text of a notice screenshot without retyping, or turning a photographed page of a book into a quotable passage.
Photographing a foreign menu or sign, extracting the text, and pasting it into a translator is another common use.
Images that recognize well
Crisp, upright text with strong contrast against the background works best. Rotate tilted photos in the Image Editor first, and raise brightness and contrast on dark shots; cropping to just the small text you need also helps.
Handwriting, decorative fonts, vertical text, and text inside tables raise the error rate. Compare the result with the original before using it.
Languages and processing
Choose Korean+English for Korean documents with some English, and plain English for English-only text, for the best accuracy. Language data is downloaded on first selection and cached by the browser, so later runs start immediately.
For multi-page scanned PDFs, the Scanned PDF OCR tool processes page by page.
Frequently asked questions
Are files uploaded to a server?+
No. Recognition happens inside your browser and images never leave your device.
Which languages are supported?+
Korean+English, English, Japanese, and Simplified Chinese.
How accurate is it?+
High for crisp printed text; handwriting, decorative fonts, and tilted photos raise errors. Compare with the original.
Why the wait the first time?+
Language data is being downloaded. It is cached afterwards, so later runs start immediately.
How can I improve accuracy?+
Rotate tilted photos, brighten dark ones, and crop to the small text you need.
Related tools