OCR (Image to Text, Japanese Support)
An OCR tool that reads Japanese and English text in images and screenshots and turns it into text, entirely in your browser. It supports selecting an area, preprocessing such as grayscale, binarization, and upscaling, and removing stray spaces in Japanese text. Your image is never sent to a server.
Drag and drop an image, or paste with Ctrl/⌘+V
Supports PNG / JPEG / WebP / BMP
How to use
- Drag and drop an image, or load one with "Choose image". You can also paste a screenshot with Ctrl/⌘+V.
- Choose the language. Use "Japanese + English" for mixed text and "Japanese (vertical)" for vertical writing.
- Adjust preprocessing (grayscale, contrast, binarization, upscaling, rotation) if needed. Drag on the preview to recognize only that area.
- Click "Recognize text" to see the result in the text area. The recognition model is downloaded only the first time.
- The result is editable. Save it with "Copy to clipboard" or "Download .txt".
About this tool
The OCR tool reads text in images and converts it into editable text. It's handy for pulling text out of photos of printed documents, screenshots, or PDFs saved as images without retyping it.
Recognition uses tesseract.js (Apache-2.0), a WebAssembly build of the open-source Tesseract OCR engine, running in a Web Worker so the page stays responsive. Because Japanese results often contain stray spaces between characters, post-processing options are included to remove them and to join line breaks within paragraphs.
The recognition engine and language data are served from this site, not from an external CDN. Language data is downloaded only once and saved in your browser, so later runs start right away. Your image never leaves your browser.
Frequently asked questions
How can I improve accuracy?
Try "Upscale" for small text and "Binarize" when the background has colors or patterns. Rotate the image to the correct orientation, and select only the area you need. Handwriting and decorative fonts are difficult to recognize.
How much data is downloaded?
The first time only, the recognition engine (about 3.9 MB) and language data (about 2 MB for Japanese or vertical Japanese, about 2.9 MB for English) are downloaded. Language data is saved in your browser and isn't downloaded again.
Can I read text from a PDF?
PDFs can't be loaded directly. Take a screenshot of the page or otherwise convert it to an image first.
Is my image sent to a server?
No. Only the recognition model is downloaded; everything from loading the image to recognizing text happens in your browser.