How to extract text from an image
- Drop one or more images onto the box above, or tap to browse.
- Choose the language of the text (tick Also read English for mixed text) and, if needed, the layout.
- Wait for OCR to process. On first use the language model downloads once — it's cached after that.
- The extracted text appears below each image and is editable. Fix any mistakes, join broken lines, then copy it or download it as a text file.
What works well
- Screenshots — highest accuracy because text is sharp and clean
- Scanned documents — with a flat, well-lit scan
- Photos of printed pages — books, receipts, forms, business cards
- Signs and labels — product labels, packaging, street signs
- Handwriting — works for neat block letters, less reliable for cursive
Tips for best accuracy
- Sharpen the image — blur and motion reduce accuracy significantly
- Crop to the text — remove surrounding noise before uploading
- Avoid shadows and glare — even lighting helps a lot
- Use the right language — and tick “Also read English” for mixed-language text
- Pick a layout — “Sparse text” suits screenshots and receipts, “Single block” suits paragraphs
- Keep Enhance image on — it evens out shadows, boosts contrast and enlarges small text
- Rotate to upright — OCR handles slight angles but not upside-down text
Your images never leave your device
The OCR engine (Tesseract) runs entirely in your browser using WebAssembly. No image, no text, no metadata is ever sent to any server. Once the language model has been downloaded, the tool works completely offline.
Why the first run is slower
The language model is between 10 and 30 MB depending on the language. It downloads once, is cached by your browser, and reused on every subsequent visit. English is the smallest (~12 MB), Chinese and Japanese the largest (~30 MB).
Common questions
Is this image to text tool free?
Yes. Completely free, no daily limit, no signup.
Does my image get uploaded?
No. The OCR engine runs on your device using WebAssembly. Your image never leaves your browser.
What languages are supported?
English, Hindi, Bengali, Tamil, Telugu, Marathi, Gujarati, Kannada, Punjabi, Urdu, Arabic, Spanish, French, German, Chinese Simplified, Japanese and Russian.
Why is the first run slower?
The OCR language model (10–30 MB) downloads once and is cached by your browser. Every image after that is much faster.
Can I extract text from a screenshot?
Yes. Screenshots usually give the best accuracy because they have sharp, clean text.
Does it read handwriting?
It handles neat block letters reasonably well. Cursive and flowing handwriting is harder and often produces errors.
Can I extract text from a PDF?
Convert the PDF to images first using our PDF to JPG tool, then upload the images here.
Is there a file size limit?
No hard limit, but very large images (over 20 MB) may take longer. Images are automatically downscaled to the optimal size for OCR.