[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-72410":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":25,"hasPages":23,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":47,"readmeContent":48,"aiSummary":49,"trendingCount":16,"starSnapshotCount":16,"syncStatus":50,"lastSyncTime":51,"discoverSource":52},72410,"comic-translate","ogkalu2\u002Fcomic-translate","ogkalu2","AI comic and manga translator app\u002Fbrowser extension for automatically translating comics, manga, manhwa, BDs, fumetti, and more in multiple languages and formats (Images, PDF, EPUB, CBR, CBZ etc).","https:\u002F\u002Fcomic-translate.com",null,"Python",2776,305,27,94,0,4,15,70,12,81.96,"Apache License 2.0",false,"main",true,[27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46],"anime","comics","computer-vision","deep-learning","gui","inpainting","machine-translation","manga","manhua","manhwa","neural-network","ocr","pyside6","python","pytorch","segmentation","text-detection","text-segmentation","translation","webtoons","2026-06-12 04:01:05","# Comic Translate\nEnglish | [한국어](docs\u002FREADME_ko.md) | [Français](docs\u002FREADME_fr.md) | [简体中文](docs\u002FREADME_zh-CN.md)\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FQUVK6mK.png\">\n\n## Intro\nMany Automatic Manga Translators exist. Very few properly support comics of other kinds in other languages. \nThis project was created to utilize the ability of State of the Art (SOTA) Large Language Models (LLMs) like GPT and translate comics from all over the world. \n\nCurrently, it supports translating comics from the following languages: English, Korean, Japanese, French, Simplified Chinese, Traditional Chinese, Russian, German, Dutch, Spanish and Italian. It can translate to the above mentioned and more. \n\n- [The State of Machine Translation](#the-state-of-machine-translation)\n- [Preview](#comic-samples)\n- [Getting Started](#installation)\n    - [Installation](#installation)\n        - [Download](#download)\n        - [From Source](#from-source)\n    - [Usage](#usage)\n        - [Tips](#tips)\n\n- [How it works](#how-it-works)\n    - [Text Detection](#text-detection)\n    - [OCR](#OCR)\n    - [Inpainting](#inpainting)\n    - [Translation](#translation)\n    - [Text Rendering](#text-rendering)\n\n- [Acknowledgements](#acknowledgements)\n\n## The State of Machine Translation\nFor a couple dozen languages, the best Machine Translator is not Google Translate, Papago or even DeepL, but a SOTA LLM like GPT-4, and by far. \nThis is very apparent for distant language pairs (Korean\u003C->English, Japanese\u003C->English etc) where other translators still often devolve into gibberish.\nExcerpt from \"The Walking Practice\"(보행 연습) by Dolki Min(돌기민)\n![Model](https:\u002F\u002Fi.imgur.com\u002F72jvLBa.png)\n\n## Comic Samples\nGPT-4 as Translator.\nNote: Some of these also have Official English Translations\n\n[The Wretched of the High Seas](https:\u002F\u002Fwww.drakoo.fr\u002Fbd\u002Fdrakoo\u002Fles_damnes_du_grand_large\u002Fles_damnes_du_grand_large_-_histoire_complete\u002F9782382330128)\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002F75HwK4r.jpg\" width=\"49%\"> \u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002F3oRt5fX.jpg\" width=\"49%\">\n\n[Journey to the West](https:\u002F\u002Fac.qq.com\u002FComic\u002FcomicInfo\u002Fid\u002F541812)\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002Fzk7yiKe.jpg\" width=\"49%\"> \u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002F4ycSi8j.jpg\" width=\"49%\">\n\n[The Wormworld Saga](https:\u002F\u002Fwormworldsaga.com\u002Findex.php)\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FcVVGVXp.jpg\" width=\"49%\"> \u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FSSl81sb.jpg\" width=\"49%\">\n\n[Frieren: Beyond Journey's End](https:\u002F\u002Frenta.papy.co.jp\u002Frenta\u002Fsc\u002Ffrm\u002Fitem\u002F220775\u002Ftitle\u002F742932\u002F)\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FANGHVhG.png\" width=\"49%\"> \u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002Fr5YOE26.png\" width=\"49%\">\n\n[Days of Sand](https:\u002F\u002F9ekunst.nl\u002F2021\u002F05\u002F20\u002Fnieuw-album-van-aimee-de-jongh-is-benauwend-als-een-zandstorm\u002F)\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002Fm7PDiXN.jpg\" width=\"49%\"> \u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FeUwTGnn.jpg\" width=\"49%\">\n\n[Player (OH Hyeon-Jun)](https:\u002F\u002Fcomic.naver.com\u002Fwebtoon\u002Flist?titleId=745876&page=1&sort=ASC&tab=fri)\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FKGwiHJh.jpg\" width=\"49%\"> \u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FB8RMbRQ.jpg\" width=\"49%\">\n\n[Carbon & Silicon](https:\u002F\u002Fwww.amazon.com\u002FCarbone-Silicium-French-Mathieu-Bablet-ebook\u002Fdp\u002FB0C1LTGZ85\u002F)\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002Fh51XJx4.jpg\" width=\"49%\"> \u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FsLitjUY.jpg\" width=\"49%\">\n\n## Installation\n### Download\nDownload and install Comic Translate for Windows and macOS from [here](https:\u002F\u002Fwww.comic-translate.com). \n\n>Ignore Smart Screen for Windows (Click More info > Run anyway). For macOS, after trying to open, go to Settings > Privacy and Security > Scroll down and click Open Anyway. \n\n> Note: GPU acceleration is currently only available when running from source.\n\n### From Source\nAlternatively, if you'd like to run the source code directly.\n\nInstall Python 3.12. Tick \"Add python.exe to PATH\" during the setup.\n```bash\nhttps:\u002F\u002Fwww.python.org\u002Fdownloads\u002F\n```\nInstall git\n```bash\nhttps:\u002F\u002Fgit-scm.com\u002F\n```\nInstall uv\n```\nhttps:\u002F\u002Fdocs.astral.sh\u002Fuv\u002Fgetting-started\u002Finstallation\u002F\n```\n\nThen, in the command line\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fogkalu2\u002Fcomic-translate\ncd comic-translate\nuv init --python 3.12\n```\nand install the requirements\n```bash\nuv add -r requirements.txt --compile-bytecode\n```\n\nTo Update, run this in the comic-translate folder\n```bash\ngit pull\nuv init --python 3.12 (Note: only run this line if you did not use uv for the first time installation)\nuv add -r requirements.txt --compile-bytecode\n```\n\nIf you have an NVIDIA GPU, then it is recommended to run\n```bash\nuv pip install onnxruntime-gpu\n```\n\n## Usage\nIn the comic-translate directory, run\n```bash\nuv run comic.py\n```\nThis will launch the GUI\n\n### Tips\n* If you have a CBR file, you'll need to install Winrar or 7-Zip then add the folder it's installed to (e.g \"C:\\Program Files\\WinRAR\" for Windows) to Path. If it's installed but not to Path, you may get the error, \n```bash\nraise RarCannotExec(\"Cannot find working tool\")\n```\nIn that case, Instructions for [Windows](https:\u002F\u002Fwww.windowsdigitals.com\u002Fadd-folder-to-path-environment-variable-in-windows-11-10\u002F), [Linux](https:\u002F\u002Flinuxize.com\u002Fpost\u002Fhow-to-add-directory-to-path-in-linux\u002F), [Mac](https:\u002F\u002Ftechpp.com\u002F2021\u002F09\u002F08\u002Fset-path-variable-in-macos-guide\u002F)\n\n* Make sure the selected Font supports characters of the target language\n* v2.0 introduces a Manual Mode. When you run into issues with Automatic Mode (No text detected, Incorrect OCR, Insufficient Cleaning etc), you are now able to make corrections. Simply Undo the Image and toggle Manual Mode.\n* In Automatic Mode, Once an Image has been processed, it is loaded in the Viewer or stored to be loaded on switch so you can keep reading in the app as the other Images are being translated.\n* Ctrl + Mouse Wheel to Zoom otherwise Vertical Scrolling\n* The Usual Trackpad Gestures work for viewing the Image\n* Right, Left Keys to Navigate Between Images\n\n## How it works\n### Speech Bubble Detection and Text Segmentation\n[bubble-and-text-detector](https:\u002F\u002Fhuggingface.co\u002Fogkalu\u002Fcomic-text-and-bubble-detector). RT-DETR-v2 model trained on 11k images of comics (Manga, Webtoons, Western).\nAlgorithmic segmentation based on the boxes provided from the detection model.\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FTlzVH3j.jpg\" width=\"49%\"> \u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002Fh18XrYT.jpg\" width=\"49%\"> \n\n### OCR\nBy Default:\n* [manga-ocr](https:\u002F\u002Fgithub.com\u002Fkha-white\u002Fmanga-ocr) for Japanese\n* [Pororo](https:\u002F\u002Fgithub.com\u002Fyunwoong7\u002Fkorean_ocr_using_pororo) for Korean \n* [PPOCRv5](https:\u002F\u002Fwww.paddleocr.ai\u002Fmain\u002Fen\u002Fversion3.x\u002Falgorithm\u002FPP-OCRv5\u002FPP-OCRv5.html) for Everything Else\n\nOptional:\n\nThese can be used for any of the supported languages.\n\n* Gemini 2.0 Flash\n* Microsoft Azure Vision\n\n### Inpainting\nTo remove the segmented text\n* A [Manga\u002FAnime finetuned](https:\u002F\u002Fhuggingface.co\u002FdreMaz\u002FAnimeMangaInpainting) [lama](https:\u002F\u002Fgithub.com\u002Fadvimman\u002Flama) checkpoint. Implementation courtsey of [lama-cleaner](https:\u002F\u002Fgithub.com\u002FSanster\u002Flama-cleaner)\n* [AOT-GAN](https:\u002F\u002Farxiv.org\u002Fabs\u002F2104.01431) based model by [zyddnys](https:\u002F\u002Fgithub.com\u002Fzyddnys)\n\n\u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FcVVGVXp.jpg\" width=\"49%\"> \u003Cimg src=\"https:\u002F\u002Fi.imgur.com\u002FbLkPyqG.jpg\" width=\"49%\">\n\n### Translation\nCurrently, this supports using GPT-4.1, Claude-4.5, \nGemini-2.5.\n\nAll LLMs are fed the entire page text to aid translations. \nThere is also the Option to provide the Image itself for further context. \n\n### Text Rendering\nWrapped text in bounding boxes obtained from bubbles and text.\n\n## Acknowledgements\n\n* [https:\u002F\u002Fgithub.com\u002FSanster\u002Flama-cleaner](https:\u002F\u002Fgithub.com\u002FSanster\u002Flama-cleaner)\n* [https:\u002F\u002Fhuggingface.co\u002FdreMaz](https:\u002F\u002Fhuggingface.co\u002FdreMaz)\n* [https:\u002F\u002Fgithub.com\u002Fyunwoong7\u002Fkorean_ocr_using_pororo](https:\u002F\u002Fgithub.com\u002Fyunwoong7\u002Fkorean_ocr_using_pororo)\n* [https:\u002F\u002Fgithub.com\u002Fkha-white\u002Fmanga-ocr](https:\u002F\u002Fgithub.com\u002Fkha-white\u002Fmanga-ocr)\n* [https:\u002F\u002Fgithub.com\u002FJaidedAI\u002FEasyOCR](https:\u002F\u002Fgithub.com\u002FJaidedAI\u002FEasyOCR)\n* [https:\u002F\u002Fgithub.com\u002FPaddlePaddle\u002FPaddleOCR](https:\u002F\u002Fgithub.com\u002FPaddlePaddle\u002FPaddleOCR)\n* [https:\u002F\u002Fgithub.com\u002FRapidAI\u002FRapidOCR](https:\u002F\u002Fgithub.com\u002FRapidAI\u002FRapidOCR)\n* [https:\u002F\u002Fgithub.com\u002Fphenom-films\u002Fdayu_widgets](https:\u002F\u002Fgithub.com\u002Fphenom-films\u002Fdayu_widgets)","comic-translate 是一个利用先进大语言模型自动翻译漫画的应用，支持多种格式如图片、PDF、EPUB、CBR 和 CBZ 等。该项目通过结合文本检测、OCR 识别、图像修复、机器翻译和文本渲染等技术，实现从包括英语、韩语、日语、法语、简体中文、繁体中文、俄语、德语、荷兰语、西班牙语和意大利语在内的多种语言到其他语言的高质量翻译。特别适用于需要跨语言阅读漫画、漫画书、网络漫画等多媒体内容的场景，尤其在处理远距离语言对（例如韩语-英语、日语-英语）时表现出色，翻译质量优于传统翻译工具。",2,"2026-06-11 03:41:55","high_star"]