Image to Text — Chinese and English OCR
Recognize Chinese and English in screenshots or scanned images, then copy text or download TXT locally.
Processed locally in your browser. Nothing is uploaded.
How to use
- Choose a static image and language.
- Load the model on first use and recognize text.
- Review, edit and copy or download text.
Example and rules
Example: recognize a clear Chinese screenshot and copy the result into your notes.
- One JPG, PNG or WebP up to 20 MiB and 20 million pixels.
- Images over 3,000 px on the longest edge are resized; only models are cached.
- PDF OCR, table reconstruction and accuracy guarantees are not included.
Frequently asked questions
Why is the first run slower?
WASM and language models must be loaded from this site.
Does it recognize handwriting?
Handwriting is not guaranteed; clear printed text works best.
Are photos uploaded?
No. Only models are downloaded; recognition is local.