About this tool
This OCR tool "reads" the text in images, screenshots and photos and turns it into text you can copy, edit and search — Chinese and English supported. No more retyping word by word; snap a photo or take a screenshot, recognize it, and copy the text out.
Recognition runs entirely in your browser on the open-source tesseract.js engine, so images are never uploaded — fine for IDs, contracts and private screenshots. First use downloads the engine and language data (a few MB), which is then cached.
How to use
- Choose the language (use "Chinese + English" for mixed text).
- Click or drag to select an image containing text.
- Wait for recognition (the first run downloads language data and is slower).
- Review and edit the result on the right, then copy or download the text.
When to use it
- Extract text from a photo of a paper document or book.
- Recognize text in screenshots or images that can’t be copied directly.
- Pull information from invoices, receipts and business cards.
- Turn a scan into editable, searchable text.
FAQ
What if recognition is inaccurate?
Accuracy depends on image clarity, contrast and font. Use clear, upright, high-contrast images; handwriting, decorative fonts, blurry or skewed text recognize less well.
Are images uploaded?
No. Recognition happens locally in your browser; images are not sent to any server.
Why is the first run slow?
The first run downloads the engine and the chosen language data (a few MB). It is cached afterward, so later recognitions are much faster.