Screen OCR Translator by Alex

Capture and translate text from images, comics, scans, or any selected area of a web page, without leaving the current tab.

Text recognition runs locally in your browser. Translation uses Google Translate by default, or you can connect it to your own LLM endpoint for translation.

How to use:

1. Open the extension from the browser toolbar or context menu, or press Ctrl+Shift+F.
2. Select an area containing text.
3. Adjust the selection if needed, then click “Run OCR.”
4. Translation appears over the original text.

To translate a single image, right-click it and select “Translate this image.”

Features:

• Local text recognition powered by PaddleOCR
• Support for Latin, Chinese, Japanese, Cyrillic, Korean, Arabic, and Devanagari scripts
• Automatic detection of the source language
• Detection of speech bubbles in manga and comics
• Translation can be displayed directly over the original content or in a separate panel
• Copy recognized or translated text to the clipboard
• Text-to-speech through Google TTS
• Connect your own LLM endpoint (llama.cpp, LM Studio, and other compatible providers)
• Support for GPU acceleration (can be enabled in settings).