圖片 OCR
使用裝置端 OCR 從圖片中擷取文字（支援英語和日語）。圖片絕不離開你的裝置；僅在首次使用時下載 OCR 引擎。
使用方法
- Choose an image that contains the text you want.
- Select the recognition language: English, Japanese, or both.
- Press "Recognize text". The first run takes a little longer while the OCR engine and language data download. You can edit the result in place, copy it, or download it as a .txt file.
注意事項
Recognition runs with Tesseract.js inside a Web Worker in your browser. Your image is never uploaded to any server. One honest caveat: on the first run only, the OCR engine and the language data are downloaded from the Tesseract CDN (jsDelivr). They are cached by your browser afterwards, so later runs start fast. The recognized text never leaves this page either.
常見問題
- How can I improve accuracy?
- Use images with large, high-contrast text. Straighten tilted photos and avoid shadows and glare. Handwriting is poorly supported and results are not guaranteed.
- Does it work offline?
- Once the engine and language data are cached by your browser it usually works offline. Clearing the browser cache means they will be downloaded again.