[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-96017":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":9,"totalLinesOfCode":9,"stars":12,"forks":13,"watchers":14,"openIssues":14,"contributorsCount":9,"subscribersCount":14,"size":14,"stars1d":14,"stars7d":14,"stars30d":14,"stars90d":14,"forks30d":14,"starsTrendScore":14,"compositeScore":15,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":16,"fork":16,"defaultBranch":17,"hasWiki":16,"hasPages":16,"topics":9,"createdAt":9,"pushedAt":9,"updatedAt":18,"readmeContent":19,"aiSummary":20,"trendingCount":14,"starSnapshotCount":14,"syncStatus":21,"lastSyncTime":9,"discoverSource":22},96017,"koharu","koharu-rs\u002Fkoharu","koharu-rs","AI-powered manga translator, written in Rust.",null,"https:\u002F\u002Fgithub.com\u002Fkoharu-rs\u002Fkoharu","Rust",5507,377,0,60.73,false,"main","2026-09-20 04:01:32","\u003Ch1 align=\"center\">Koharu\u003C\u002Fh1>\n\n\u003Cp align=\"center\">ML-powered manga translator, written in \u003Cb>Rust\u003C\u002Fb>.\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fkoharu-rs\u002Fkoharu\u002Freleases\u002Flatest\" target=\"_blank\">\u003Cimg alt=\"GitHub Downloads (all assets, all releases)\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fdownloads\u002Fkoharu-rs\u002Fkoharu\u002Ftotal?style=for-the-badge&link=https%3A%2F%2Fgithub.com%2Fkoharu-rs%2Fkoharu%2Freleases%2Flatest\">\u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n\u003Ca href=\"https:\u002F\u002Ftrendshift.io\u002Frepositories\u002F20649\" target=\"_blank\">\u003Cimg src=\"https:\u002F\u002Ftrendshift.io\u002Fapi\u002Fbadge\u002Frepositories\u002F20649\" alt=\"koharu-rs%2Fkoharu | Trendshift\" style=\"width: 250px; height: 55px;\" width=\"250\" height=\"55\"\u002F>\u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n\u003Ca href=\"https:\u002F\u002Fkoharu.rs\u002Fen\u002Finstallation\" target=\"_blank\">Getting Started\u003C\u002Fa> · \u003Ca href=\"https:\u002F\u002Fkoharu.rs\u002F\" target=\"_blank\">Docs\u003C\u002Fa> · \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fkoharu-rs\u002Fkoharu\u002Fissues\" target=\"_blank\">Bug reports\u003C\u002Fa> · \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FmHvHkxGnUY\" target=\"_blank\">Discord\u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n\u003Ca href=\"https:\u002F\u002Fkoharu.rs\u002Fja\" target=\"_blank\">日本語\u003C\u002Fa> | \u003Ca href=\"https:\u002F\u002Fkoharu.rs\u002Fzh\" target=\"_blank\">简体中文\u003C\u002Fa>\n\u003C\u002Fp>\n\nKoharu introduces a local-first workflow for manga translation, utilizing the power of ML to automate the process. It combines the capabilities of object detection, OCR, inpainting, and LLMs to create a seamless translation experience.\n\n> [!NOTE]\n> Koharu runs its vision models and LLMs **locally** on your machine to keep your data private and secure.\n\n---\n\n![screenshot](packages\u002Fdocs\u002Fscreenshot.png)\n\n> [!NOTE]\n> Join our [Discord server](https:\u002F\u002Fdiscord.gg\u002FmHvHkxGnUY) for support and discussion.\n\n## Features\n\n- [Multi-format project management](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fguides\u002Fprojects) for raster images, archives, and PDFs with page sequencing\n- [Selective pipeline](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fguides\u002Fprocessing) for detection, OCR, translation, and inpainting at page or project scope\n- [Detection and segmentation](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fguides\u002Fprocessing) for text regions, speech bubbles, and cleanup regions\n- [Multimodal OCR](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fmodels\u002Fvision) for dialogue, captions, and general page text\n- [Local GGUF inference and hosted providers](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fmodels\u002Fproviders) for LLM and machine-translation workflows\n- [Generative inpainting](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fguides\u002Fcleanup) for source-text removal and artwork reconstruction\n- [Proofreading](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fguides\u002Freview) for correcting OCR and translation output\n- [WebGPU-based canvas](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fguides\u002Fcanvas) for manual cleanup, text placement, and page composition\n- [Multilingual text shaping and layout](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fguides\u002Ftypesetting) with automatic fitting, font fallback, vertical CJK, and right-to-left text\n- [Layered PSD export](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fguides\u002Fexport) for flattened delivery and layered editing\n- [Agent-based workflow](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fagent\u002Fprojects) for project inspection, editing, and pipeline control\n\n## Hardware Acceleration\n\nKoharu supports CUDA and ROCm \u002F HIP on Windows and Linux, Metal on Apple silicon, and Vulkan on Windows and Linux. Keep your graphics driver current; a full CUDA or ROCm SDK installation is not required. See [Runtime and hardware requirements](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fhardware) for model-specific guidance.\n\n### CUDA\n\nCUDA 13.3 requires an NVIDIA Turing-class or newer GPU and an R610 or newer driver. Check NVIDIA's official [CUDA toolkit, driver, and architecture matrix](https:\u002F\u002Fdocs.nvidia.com\u002Fdatacenter\u002Ftesla\u002Fdrivers\u002Fcuda-toolkit-driver-and-architecture-matrix.html) and install the [latest NVIDIA driver](https:\u002F\u002Fwww.nvidia.com\u002Fen-us\u002Fdrivers\u002F).\n\n### ROCm \u002F HIP\n\nROCm 10.0 support depends on the exact AMD GPU, operating system, and driver combination. Check AMD's official [ROCm 10.0.0 compatibility matrix](https:\u002F\u002Frocm.docs.amd.com\u002Fen\u002Fdocs-10.0.0\u002Fcompatibility\u002Fcompatibility-matrix.html) and install a compatible [AMD driver](https:\u002F\u002Fwww.amd.com\u002Fen\u002Fsupport).\n\n### Metal\n\nMetal is available on Apple silicon Macs.\n\n### Vulkan\n\nVulkan is available on Windows and Linux as an alternative to CUDA and ROCm \u002F HIP.\n\n### WebGPU\n\nThe editor canvas uses WebGPU and requires a current graphics driver even when inference runs on the CPU.\n\n### CPU\n\nCPU inference is available for supported workloads but is substantially slower.\n\n## Machine Learning Models\n\nKoharu uses separate models for detection, OCR, inpainting, and translation. [Vision and inpainting](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fmodels\u002Fvision) and [translation and generation](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fmodels\u002Ftranslation) have separate model settings.\n\n### Computer Vision Models\n\nDetection, OCR, and inpainting models are selected separately.\n\n#### Detection and Layout\n\nThe detection model finds text regions, speech bubbles, and segmentation masks.\n\n- [Koharu Layout RF-DETR Seg 2XL](https:\u002F\u002Fhuggingface.co\u002Fmayocream\u002Fkoharu-layout-rfdetr-seg-2xl-1152)\n\n#### OCR\n\nOCR reads source text from detected regions.\n\n- [PaddleOCR VL 1.6](https:\u002F\u002Fhuggingface.co\u002FPaddlePaddle\u002FPaddleOCR-VL-1.6)\n- [Manga OCR](https:\u002F\u002Fhuggingface.co\u002Fmayocream\u002Fmanga-ocr)\n- [Baberu OCR](https:\u002F\u002Fhuggingface.co\u002Fgenshiai-daichi\u002Fbaberu-ocr)\n- [Hayai OCR](https:\u002F\u002Fhuggingface.co\u002FJustANormalTinkerer\u002Fhayai-ocr-v2)\n\n#### Inpainting\n\nInpainting reconstructs the image behind source text before the translation is rendered.\n\n- [FLUX.2 Klein](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FFLUX.2-klein-4B-GGUF)\n- [RORem mixed](https:\u002F\u002Fhuggingface.co\u002Fmayocream\u002FRORem-mixed-GGUF)\n- [LaMa](https:\u002F\u002Fhuggingface.co\u002Fmayocream\u002Flama-manga)\n- [AOT GAN](https:\u002F\u002Fhuggingface.co\u002Fmayocream\u002Faot-inpainting)\n\n### Large Language Models\n\nTranslation can use a local language model or a remote API.\n\n#### General-Purpose Local Models\n\n- LFM 2.5: [lfm2.5-1.2b-instruct](https:\u002F\u002Fhuggingface.co\u002FLiquidAI\u002FLFM2.5-1.2B-Instruct-GGUF)\n- Ministral 3: [ministral-3-8b-instruct](https:\u002F\u002Fhuggingface.co\u002Fmistralai\u002FMinistral-3-8B-Instruct-2512-GGUF)\n- Gemma 4: [gemma4-e2b-it](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002Fgemma-4-E2B-it-qat-GGUF), [gemma4-e4b-it](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002Fgemma-4-E4B-it-qat-GGUF), [gemma4-12b-it](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002Fgemma-4-12B-it-qat-GGUF), [gemma4-26b-a4b-it](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002Fgemma-4-26B-A4B-it-qat-GGUF), [gemma4-31b-it](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002Fgemma-4-31B-it-qat-GGUF)\n- Qwen 3.5: [qwen3.5-0.8b](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FQwen3.5-0.8B-GGUF), [qwen3.5-2b](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FQwen3.5-2B-GGUF), [qwen3.5-4b](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FQwen3.5-4B-GGUF), [qwen3.5-9b](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FQwen3.5-9B-GGUF), [qwen3.5-27b](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FQwen3.5-27B-GGUF), [qwen3.5-35b-a3b](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FQwen3.5-35B-A3B-GGUF)\n- Qwen 3.6: [qwen3.6-27b](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FQwen3.6-27B-GGUF), [qwen3.6-35b-a3b](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FQwen3.6-35B-A3B-GGUF)\n- Qwen 3.8: [qwen3.8-27b](https:\u002F\u002Fhuggingface.co\u002Funsloth\u002FQwen3.8-27B-GGUF)\n\n#### Uncensored Local Models\n\n- Gemma 4 uncensored: [gemma4-e2b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FGemma-4-E2B-Uncensored-HauhauCS-Aggressive), [gemma4-e4b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FGemma-4-E4B-Uncensored-HauhauCS-Aggressive), [gemma4-12b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FGemma4-12B-QAT-Uncensored-HauhauCS-Balanced), [gemma4-26b-a4b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FGemma4-26B-A4B-QAT-Uncensored-HauhauCS-Balanced-MTP), [gemma4-31b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FGemma4-31B-QAT-Uncensored-HauhauCS-Balanced-MTP)\n- Qwen 3.5 uncensored: [qwen3.5-2b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FQwen3.5-2B-Uncensored-HauhauCS-Aggressive), [qwen3.5-4b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FQwen3.5-4B-Uncensored-HauhauCS-Aggressive), [qwen3.5-9b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FQwen3.5-9B-Uncensored-HauhauCS-Aggressive)\n- Qwen 3.6 uncensored: [qwen3.6-27b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FQwen3.6-27B-Uncensored-HauhauCS-Balanced), [qwen3.6-35b-a3b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FQwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive)\n- Qwen 3.8 uncensored: [qwen3.8-27b-uncensored](https:\u002F\u002Fhuggingface.co\u002FHauhauCS\u002FQwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF)\n\n#### Cloud Providers\n\nHosted LLM providers: [Atlas Cloud](https:\u002F\u002Fwww.atlascloud.ai\u002F), [OpenAI](https:\u002F\u002Fplatform.openai.com\u002F), [Gemini](https:\u002F\u002Fai.google.dev\u002F), [Claude](https:\u002F\u002Fwww.anthropic.com\u002Fapi), [Grok](https:\u002F\u002Fdocs.x.ai\u002Fdevelopers), [MiniMax](https:\u002F\u002Fplatform.minimax.io\u002F), [DeepSeek](https:\u002F\u002Fplatform.deepseek.com\u002F), and [OpenRouter](https:\u002F\u002Fopenrouter.ai\u002F).\n\n#### Machine Translation Providers\n\nMachine-translation providers: [DeepL](https:\u002F\u002Fwww.deepl.com\u002F), [Google Cloud Translation](https:\u002F\u002Fcloud.google.com\u002Ftranslate), and [Caiyun](https:\u002F\u002Ffanyi.caiyunapp.com\u002F).\n\n#### OpenAI-Compatible Providers\n\nOpenAI-compatible endpoints are also supported.\n\n## Installation\n\nDownload release builds from the [releases page](https:\u002F\u002Fgithub.com\u002Fkoharu-rs\u002Fkoharu\u002Freleases\u002Flatest). [Installation requirements and first launch](https:\u002F\u002Fkoharu.rs\u002Fen\u002Finstallation) vary by operating system.\n\nBuilds are available for Windows, macOS, and Linux.\n\n### WinGet\n\nInstall on Windows with [winget](https:\u002F\u002Flearn.microsoft.com\u002Fen-us\u002Fwindows\u002Fpackage-manager\u002Fwinget\u002F):\n\n```bash\nwinget install koharu\n```\n\n### Homebrew\n\nInstall on macOS with [Homebrew](https:\u002F\u002Fbrew.sh\u002F):\n\n```bash\nbrew install --cask koharu\n```\n\n## Troubleshooting\n\nStartup, runtime, model, and provider errors are covered in [Troubleshooting](https:\u002F\u002Fkoharu.rs\u002Fen\u002Freference\u002Ftroubleshooting). Set `RUST_LOG` to `debug` or `trace` for verbose logs:\n\n```bash\n# macOS \u002F Linux\nRUST_LOG=debug koharu\n# Windows (PowerShell)\n$env:RUST_LOG=\"debug\"; koharu.exe\n```\n\n## Development\n\nPlatform dependencies and validation commands for local builds are listed in [Development Setup](https:\u002F\u002Fkoharu.rs\u002Fen\u002Fdevelopment\u002Fsetup).\n\n### Prerequisites\n\n- [Rust](https:\u002F\u002Fwww.rust-lang.org\u002Ftools\u002Finstall) 1.97.1 or later (Rust 2024 edition)\n- [Bun](https:\u002F\u002Fbun.sh\u002F) 1.3.14 or later\n- [LLVM](https:\u002F\u002Fllvm.org\u002F) 22.1.8 or later\n\n### Install dependencies\n\n```bash\nbun install\n```\n\n### Development\n\n```bash\nbun dev\n```\n\n### Build\n\n```bash\nbun run build\n```\n\nThe executable is written to `target\u002Frelease`.\n\n## Sponsorship\n\nIf Koharu is useful in your workflow, consider sponsoring the project.\n\n- [GitHub Sponsors](https:\u002F\u002Fgithub.com\u002Fsponsors\u002Fmayocream)\n- [Patreon](https:\u002F\u002Fwww.patreon.com\u002Fmayocream)\n\n![sponsors](.\u002F.github\u002Fsponsorkit\u002Fsponsors.svg)\n\n## Contributors ❤️\n\nThanks to all the contributors who have helped make Koharu better!\n\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fkoharu-rs\u002Fkoharu\u002Fgraphs\u002Fcontributors\">\n  \u003Cimg src=\"https:\u002F\u002Fcontrib.rocks\u002Fimage?repo=koharu-rs\u002Fkoharu\" \u002F>\n\u003C\u002Fa>\n\n## License\n\nCopyright 2025-2026 Mayo Takanashi and Koharu contributors.\n\nKoharu is dual-licensed under the [MIT License](LICENSE-MIT) or the\n[Apache License, Version 2.0](LICENSE-APACHE), at your option.\n","Koharu 是一款基于 Rust 开发的本地化漫画翻译工具，利用机器学习实现端到端自动化翻译流程。其核心功能涵盖文本区域检测、多模态 OCR（支持对话\u002F标题\u002F普通文本）、本地运行的 GGUF 格式 LLM 翻译、生成式图文修复（inpainting）以及多语言排版（含 CJK 垂直布局与 RTL 支持）。所有模型均默认在用户设备本地执行，保障隐私安全。适用于个人漫画汉化、小团队本地化协作及需数据离线处理的出版前翻译场景。",2,"trending"]