Budou - OCR Scanlator 作者: BudouDev
On-demand image translator and scanlation, right in your browser. Get simple translations with booru-style overlays on any image, or plug in your own translator and export finished, translated images with a full editing suite.
扩展元数据
关于此扩展
Budou translates images and text right where you're reading them — no copying text out, no switching to another tab, no uploading files anywhere by hand.
Translate any image in place
Right-click an image (manga, manhwa, comics, screenshots, memes, anything with text in it) and Budou detects the text, translates it, and overlays the translation directly on top — positioned line-for-line over the original. Boxes are fully editable afterward: move them, resize them, edit the translation text, split a box that merged two speech bubbles, or add a box by hand for anything the detector missed.
Translate selected text
Highlight any text on a page and a small icon appears for one-click translation, or just right-click → Translate. Shows inline, no page reload.
Fully offline OCR mode
By default, Budou uses Google Lens to read text out of images and Google Translate to translate it, which sends image/text data to Google's servers (see the privacy policy below for specifics). If you'd rather nothing ever leaves your device, turn on Manga OCR in Settings: a real text detector and recognizer are bundled directly into the extension and run entirely inside your browser. No server, no account, no internet connection required for the OCR step itself, and no image data is ever transmitted anywhere. Currently covers Japanese, Chinese, and Korean (including vertical text columns). Support for more scripts is in development.
Bring your own AI
Prefer to use your own local language model instead of Google Translate? Point Budou at any OpenAI-API-compatible server (Ollama, LM Studio, etc.) running on your machine or network, and translations go there instead. A "dialogue mode" toggle nudges the model to translate manga/manhwa dialogue naturally rather than stiffly.
Clean, exportable results
Export a translated image as a PNG with the overlay baked in, or go further with inpainting (via a local ComfyUI + LaMa setup, or a fully built-in OpenCV-based option) to erase the original text and composite your translation onto a clean background.
Everything else
A glossary for consistent terminology across a whole series, a translation cache for context-aware local translations, automatic continuous-scroll translation for webtoon-style reading, a screen-region capture tool for translating anything on screen, and full customization of overlay color, opacity, font, and export destination. Tweak and tune your overlay for your perfect TL.
Translate any image in place
Right-click an image (manga, manhwa, comics, screenshots, memes, anything with text in it) and Budou detects the text, translates it, and overlays the translation directly on top — positioned line-for-line over the original. Boxes are fully editable afterward: move them, resize them, edit the translation text, split a box that merged two speech bubbles, or add a box by hand for anything the detector missed.
Translate selected text
Highlight any text on a page and a small icon appears for one-click translation, or just right-click → Translate. Shows inline, no page reload.
Fully offline OCR mode
By default, Budou uses Google Lens to read text out of images and Google Translate to translate it, which sends image/text data to Google's servers (see the privacy policy below for specifics). If you'd rather nothing ever leaves your device, turn on Manga OCR in Settings: a real text detector and recognizer are bundled directly into the extension and run entirely inside your browser. No server, no account, no internet connection required for the OCR step itself, and no image data is ever transmitted anywhere. Currently covers Japanese, Chinese, and Korean (including vertical text columns). Support for more scripts is in development.
Bring your own AI
Prefer to use your own local language model instead of Google Translate? Point Budou at any OpenAI-API-compatible server (Ollama, LM Studio, etc.) running on your machine or network, and translations go there instead. A "dialogue mode" toggle nudges the model to translate manga/manhwa dialogue naturally rather than stiffly.
Clean, exportable results
Export a translated image as a PNG with the overlay baked in, or go further with inpainting (via a local ComfyUI + LaMa setup, or a fully built-in OpenCV-based option) to erase the original text and composite your translation onto a clean background.
Everything else
A glossary for consistent terminology across a whole series, a translation cache for context-aware local translations, automatic continuous-scroll translation for webtoon-style reading, a screen-region capture tool for translating anything on screen, and full customization of overlay color, opacity, font, and export destination. Tweak and tune your overlay for your perfect TL.
评分 5(1 位用户)
权限与数据
更多信息
- 版本
- 1.0.1
- 大小
- 31.46 MB
- 上次更新
- 6 天前 (2026年8月8日)
- 隐私政策
- 阅读此附加组件的隐私政策
- 版本历史
- 添加到收藏集