Free Chinese OCR, in your browser

Chinese OCR is really two problems, because the script has two character sets. Tesseract ships a separate model for each, and this converter offers both in the language dropdown above:

• Chinese (Simplified) — the default here. Mainland China, Singapore, Malaysia; most websites, apps and modern documents.
• Chinese (Traditional) — Taiwan, Hong Kong, Macau, and older printed books everywhere. Switch the dropdown before converting.

The choice matters: many characters differ between the sets (书 vs 書, 点 vs 點), and running a Traditional document through the Simplified model — or vice versa — costs real accuracy. If a conversion comes back with odd character substitutions, the wrong set is the first thing to check.

Like Japanese, Chinese is written without spaces between words, and raw Tesseract output inserts artificial segmentation spaces anyway. This converter strips a space only when both neighbours are Chinese characters, so the text reads naturally while real spacing around embedded Latin words and numbers survives. Everything runs on your device: the model downloads to your browser on first use (it's a large one, so the first conversion is the slow one) and your image never leaves your machine.

Horizontal printed text is the reliable case. Screenshots, websites, chat apps, slides and modern documents read well. Vertical columns — common in Traditional books and newspapers — are not supported and come back scrambled.

Handwriting is out of scope. The models are trained on printed type. Handwritten notes and calligraphy mostly won't survive; don't rely on them.

Dense pages need a decent capture. Chinese characters carry a lot of strokes in a small box, so resolution matters more than it does for Latin text. The built-in preprocessing (grayscale, contrast, upscaling) helps small text, but a sharp, straight-on capture beats any cleanup.

Mixed Chinese-and-Latin text mostly works. Embedded English names, numbers and dates come through with their real spacing kept.