Chinese and English together
Use screenshots, menus, signs, study notes, and scanned documents. The multilingual model handles mixed scripts in the same line.
中英文识别 · Chinese OCR
Extract printed Chinese text from images
Extract simplified Chinese, traditional Chinese, English, numbers, and mixed lines from a queue of up to 10 images—privately, with no language switch and no upload.
✓ No uploads✓ English + Chinese✓ Works offline after caching
Choose one model for the queue
Start with Fast for everyday screenshots. High accuracy requires an explicit download confirmation and reruns every queued image when selected.
Keep sensitive documents local
Identity documents and business screenshots never need to leave your device. Detection, recognition, editing, and ZIP creation stay local.
Edit before you export
Review each text or table Markdown result, then copy it, download it individually, or include every successful item in the batch ZIP.
Common questions
Frequently asked questions
Does it read simplified and traditional Chinese?
Yes. PP-OCRv6 uses a multilingual recognition dictionary covering simplified Chinese, traditional Chinese, English, numbers, and punctuation.
Can it read mixed Chinese and English?
Yes. Mixed-language lines are recognized by the same model, so you do not need to pick a language first.
Can I process several images?
Yes. Add up to 10 images or 100 MB; BrowserOCR reads them in order and lets you edit each result.
Will High accuracy find more text?
Not necessarily. It upgrades recognition after lines are detected; both modes use the same Tiny detector.
A reproducible mixed-text example
Checked: · Chrome 152.0.7977.76
Synthetic screenshot used by our OCR quality test. Arial / PingFang SC, 58 px, 1000 × 400 pixels; Chrome on macOS. This is a controlled example, not a document accuracy benchmark.
- Expected text
浏览器 OCR 安全 PRIVATE- Fast result
浏览器OCR安全PRIVATE- High accuracy result
烯OCR纬摧PRIVATE
Fast removes spaces in this sample. High accuracy makes Chinese character errors here: a larger recognizer does not guarantee a better result. Both use the Tiny detector. Review names, amounts and punctuation against the image before exporting. Blur, small print, handwriting, vertical text and complex layouts need separate checks; this sample does not establish their quality.