Chinese OCR — Extract Chinese Text From Any Image
Convert images containing Chinese into editable text. Both Simplified (简体) and Traditional (繁體) characters are recognised, and the output preserves the characters themselves rather than converting them to pinyin.
Drop your image here
or click to browse — JPG, PNG, WEBP, GIF…
Upload an image and click Extract Text
All readable text from the image will appear here.
What Makes Chinese OCR Difficult
Chinese OCR faces a scale problem no alphabetic script has: instead of 26 letters there are thousands of distinct characters in everyday use, and many differ by a single stroke. 未 and 末, 日 and 曰, or 戈 and 弋 are separated by details that survive only if the image has enough resolution. Dense characters like 麗 or 體 can collapse into a grey blob at small sizes. Chinese also writes without spaces between words, so there is no word boundary for the model to anchor on, and older or formal material may be typeset vertically, top-to-bottom and right-to-left.
Getting Accurate Chinese Results
- Resolution matters more here than for any Latin-script language. Aim for characters at least 25–30 pixels tall; complex characters need more.
- Avoid heavy JPEG compression — the artefacts it creates around fine strokes are exactly what makes 未 read as 末.
- Vertical traditional typesetting (common in older books and Taiwanese print) is read less reliably than horizontal. Crop column by column if the result looks scrambled.
- Simplified and Traditional both work without you selecting a variant, but don't expect conversion between them — you get back whichever was printed.
Who Uses Chinese OCR
Language learners
Lift characters off menus, signs and textbook pages so you can look them up, add them to flashcards or paste them into a dictionary.
Import and trade paperwork
Extract product names, specifications and company details from Chinese packaging, labels, invoices and certificates.
Research and archives
Make scanned Chinese books, newspapers and reports searchable instead of re-keying thousands of characters by hand.
Chinese OCR — Frequently Asked Questions
Does it handle both Simplified and Traditional Chinese?▼
Yes, and you don't need to choose. The model recognises whichever variant appears in the image and returns those same characters. It does not convert between Simplified and Traditional — for that, run the extracted text through a converter afterwards.
Do I get characters or pinyin?▼
Chinese characters. The output is the text as printed, in Han characters, ready to paste anywhere. No romanisation is applied.
Can it read vertical Chinese text?▼
Often, but less reliably than horizontal text. Vertical columns are most common in older books, classical literature and some Taiwanese publications. If the output looks jumbled, crop and process one column at a time.
Will it read handwritten Chinese?▼
Neat, printed-style handwriting sometimes works. Running hand (行書) and cursive (草書) are not reliably recognised — the strokes merge and abbreviate too far from printed forms. Use printed or typed source material where accuracy matters.
Need a Translation Instead?
This page extracts Chinese text as written. If you want the meaning in another language, the image translator reads the text and translates it in one step.
OCR in Other Languages
Or use the general image to text converter if your image contains several languages at once.