How to use it
Pick the language in your image, then upload or drag in a JPG, PNG or WebP. The tool uses OCR (optical character recognition) to read the text, which you can then edit, copy or download as a text file. The first scan downloads the language model once; after that it is cached and works offline.
What you can use it for
- Copy text from a screenshot or photo instead of retyping it.
- Digitise a printed notice, form, book page or receipt (English or Nepali).
- Pull a phone number, address or paragraph out of an image quickly.
Tips for the best results
- Use a sharp, well-lit image — dark text on a light background.
- Crop to just the text and keep it straight (not tilted).
- Pick the correct language; use “Both” only when the image truly mixes scripts.
- Printed/typed text works well; handwriting is unreliable.
Accuracy & privacy
OCR is an estimate, not a perfect transcription — always proofread, especially numbers and Nepali text. Everything runs locally in your browser with the open-source Tesseract.jsengine, so your image and the extracted text never leave your device.
Working with images? See the Image Compressor, Image Converter and PDF to JPG.
Frequently asked questions
How do I convert an image to text?
Choose the language in the image (English, Nepali or both), upload or drop your photo or screenshot, and the tool reads the text using OCR. Copy it or download as a .txt file. Everything runs in your browser — the image is never uploaded.
Does it work with Nepali (Devanagari) text?
Yes. Select Nepali (नेपाली) or Both. Nepali OCR works on clear printed text — notices, books, forms, screenshots — but it is less accurate than English and struggles with stylised fonts or low-quality scans. Always proofread the result.
Why is the first scan slow?
The first time, your browser downloads the OCR engine and the language model (a few MB). It is then cached, so later scans in the same browser are fast and work even offline.
How do I get the best accuracy?
Use a sharp, well-lit, straight image with good contrast (dark text on a light background). Crop to just the text, avoid shadows and angles, and pick the correct language. Printed text reads far better than handwriting.
Can it read handwriting?
Not reliably. Tesseract OCR is built for printed/typed text. Handwriting, signatures and decorative fonts often come out wrong.
Is my image private?
Yes. OCR happens entirely on your device with Tesseract.js — the image and the extracted text are never sent to any server.