Back

Image tools

Image text recognition (OCR)

Recognize printed English, Simplified Chinese, Traditional Chinese or Japanese text using local OCR models.

Free to useProcessed on your device
Open tool…

About this tool

Extract editable text from a clear image using Tesseract running in a local Web Worker. Language models and the recognition engine are served by this website; the image stays on your device.

How to use

  1. Choose a clear, upright image containing printed text. The image is processed on your device.
  2. Select the recognition language and choose automatic page, single text block or single line layout.
  3. Run recognition, proofread the extracted text, then copy it or download a UTF-8 text file.
Supported models: English, Simplified Chinese, Traditional Chinese and Japanese, plus mixed English and Simplified Chinese. Input is limited to 10 MB and 8 megapixels; images are scaled to a maximum 2400-pixel long edge. OCR can use substantial browser memory and may misread handwriting, small characters or complex layouts. Proofread the output. The engine confidence score is not a guarantee of accuracy.

Comment

How can we help?

Tell us about your tool, workflow, or partnership idea.