Extract printed text from an image with OCR
Recognize English, Arabic or mixed printed text from a JPG, PNG, WebP or BMP image. Tesseract.js runs in a background worker and all model files are served from OmniTools.
How to use Image to Text (OCR)
- Choose a clear image within the displayed limits.
- Select the text language.
- Run OCR, review the preview and download the TXT result.
Improve image OCR accuracy
Use an upright, tightly cropped image with sharp dark text on a light background. Handwriting, decorative fonts, blur and low contrast reduce accuracy.
Frequently asked questions
Does OCR work with handwriting?
It is designed for printed text and is not reliable for handwriting.
Are images uploaded?
No. Recognition runs inside your browser.
Why does the first run take longer?
The browser must load and cache the local OCR engine and selected language data.