How to use the Image to Text (OCR)
- Drop images, click to choose, or paste a screenshot with Ctrl+V; on a phone you can take a photo.
- Choose the language: English + Chinese, Simplified Chinese or English.
- To read only part of an image, drag a box on the preview.
- Click “Extract text”. Edit the result in the text box, copy it or download a TXT file.
How to get the best OCR results
Sharp, straight text works best. Fill the frame with the page, avoid glare and shadows, and shoot square-on; screenshots need no preparation. For very small text (under about 10 px tall) tick “Enhance photo”, which enlarges and boosts contrast first. For a tilted or perspective-distorted page, straighten it with the Document Scanner and then read it here.
Why does Chinese text show spaces or wrong characters?
Tesseract’s Chinese model treats every character as a word, so raw output often has spaces between characters; this tool removes them by default. Handwriting, decorative fonts, vertical text and complex tables are harder. Check look-alike characters against the image, and pay extra attention when the page warns about low confidence.
Privacy and offline use
Recognition runs in a Web Worker in your browser. The page downloads the OCR engine (about 4 MB) and the language data you pick (English about 3 MB, Chinese about 1.7 MB) from this site once and then caches them. Images and text are never uploaded, which makes it safer than online OCR services for IDs, contracts or medical papers.
Works well with PDF, translation and barcode tools
For multi-page scanned PDFs use the OCR feature in our PDF tools to process the whole file. Paste recognised foreign text into the translator. For QR codes and barcodes on receipts or packaging use the QR & barcode scanner, which is faster and exact.
Frequently asked questions
Are my images uploaded?
No. The OCR engine and language data download from this site and run in your browser. Images and results exist only in the open page and disappear when you close it.
Which languages are supported?
English and Simplified Chinese, including mixed text. Traditional Chinese, Japanese and Korean are not supported yet. The English model also reads digits and common symbols.
Can it read handwriting?
Neat, print-like handwriting sometimes works; cursive and joined-up writing usually does not. Tesseract is designed for printed text.
How many images can I process?
Add as many as you like; they are read one by one and the results are grouped by file. Each image can be up to 40 MB and 40 megapixels.
Why is the first run slower?
The first run downloads the engine and language data (about 9 MB for English + Chinese). After that the browser cache makes recognition start almost instantly.
Does it keep the layout?
Line breaks and paragraphs are kept. If a paragraph was hard-wrapped, tick “Join broken lines” for flowing text. Tables come out as space-separated text, so complex tables need manual tidying.