Image to text
Extract printed text from a still JPG, PNG or WebP using local OCR in ten supported languages. Choose the printed language, review the recognized text, and copy or download it.
Also known asimage OCR · extract text from image
OCR can misread text. Review the result before downloading.
Three small steps.
- Choose one still JPG, PNG or WebP and select its printed language.
- Start recognition; the local language model loads from this site.
- Review and correct the recognized text, then copy it or download a text file.
A few things to know.
One image up to 25 MB, 24 megapixels and 12,000 pixels per side. OCR supports English, Simplified Chinese, Hindi, Spanish, French, Arabic, Bengali, German, Italian and Portuguese, one language at a time. Indonesian and Urdu are interface languages, not available OCR models. Handwriting and exact page layout are unsupported. Recognition can misread names, numbers and punctuation; review every result. The image and text are not uploaded.
How your privacy works →A little more clarity.
Does OCR send my image to a server?
No. The engine and selected language model load from this site, but recognition happens in your browser. Your image and recognized text remain in the tab.
Can OCR make mistakes?
Yes. Small text, blur, rotation, complex backgrounds and unusual fonts can reduce accuracy. Choose the correct printed language and review the result before using it.