koharu

Koharu

ML-powered manga translator, written in Rust.

Getting Started · Docs · Bug reports · Discord

日本語 | 简体中文

Koharu 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. > [!NOTE] > Koharu runs its vision models and LLMs **locally** on your machine to keep your data private and secure. --- > [!NOTE] > Join our [Discord server](https://discord.gg/mHvHkxGnUY) for support and discussion. ## Features - [Multi-format project management](https://koharu.rs/en/guides/projects) for raster images, archives, and PDFs with page sequencing - [Selective pipeline](https://koharu.rs/en/guides/processing) for detection, OCR, translation, and inpainting at page or project scope - [Detection and segmentation](https://koharu.rs/en/guides/processing) for text regions, speech bubbles, and cleanup regions - [Multimodal OCR](https://koharu.rs/en/models/vision) for dialogue, captions, and general page text - [Local GGUF inference and hosted providers](https://koharu.rs/en/models/providers) for LLM and machine-translation workflows - [Generative inpainting](https://koharu.rs/en/guides/cleanup) for source-text removal and artwork reconstruction - [Proofreading](https://koharu.rs/en/guides/review) for correcting OCR and translation output - [WebGPU-based canvas](https://koharu.rs/en/guides/canvas) for manual cleanup, text placement, and page composition - [Multilingual text shaping and layout](https://koharu.rs/en/guides/typesetting) with automatic fitting, font fallback, vertical CJK, and right-to-left text - [Layered PSD export](https://koharu.rs/en/guides/export) for flattened delivery and layered editing - [Agent-based workflow](https://koharu.rs/en/agent/projects) for project inspection, editing, and pipeline control ## Hardware Acceleration Koharu supports CUDA and ROCm / 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://koharu.rs/en/hardware) for model-specific guidance. ### CUDA CUDA 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://docs.nvidia.com/datacenter/tesla/drivers/cuda-toolkit-driver-and-architecture-matrix.html) and install the [latest NVIDIA driver](https://www.nvidia.com/en-us/drivers/). ### ROCm / HIP ROCm 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://rocm.docs.amd.com/en/docs-10.0.0/compatibility/compatibility-matrix.html) and install a compatible [AMD driver](https://www.amd.com/en/support). ### Metal Metal is available on Apple silicon Macs. ### Vulkan Vulkan is available on Windows and Linux as an alternative to CUDA and ROCm / HIP. ### WebGPU The editor canvas uses WebGPU and requires a current graphics driver even when inference runs on the CPU. ### CPU CPU inference is available for supported workloads but is substantially slower. ## Machine Learning Models Koharu uses separate models for detection, OCR, inpainting, and translation. [Vision and inpainting](https://koharu.rs/en/models/vision) and [translation and generation](https://koharu.rs/en/models/translation) have separate model settings. ### Computer Vision Models Detection, OCR, and inpainting models are selected separately. #### Detection and Layout The detection model finds text regions, speech bubbles, and segmentation masks. - [Koharu Layout RF-DETR Seg 2XL](https://huggingface.co/mayocream/koharu-layout-rfdetr-seg-2xl-1152) #### OCR OCR reads source text from detected regions. - [PaddleOCR VL 1.6](https://huggingface.co/PaddlePaddle/PaddleOCR-VL-1.6) - [Manga OCR](https://huggingface.co/mayocream/manga-ocr) - [Baberu OCR](https://huggingface.co/genshiai-daichi/baberu-ocr) - [Hayai OCR](https://huggingface.co/JustANormalTinkerer/hayai-ocr-v2) #### Inpainting Inpainting reconstructs the image behind source text before the translation is rendered. - [FLUX.2 Klein](https://huggingface.co/unsloth/FLUX.2-klein-4B-GGUF) - [RORem mixed](https://huggingface.co/mayocream/RORem-mixed-GGUF) - [LaMa](https://huggingface.co/mayocream/lama-manga) - [AOT GAN](https://huggingface.co/mayocream/aot-inpainting) ### Large Language Models Translation can use a local language model or a remote API. #### General-Purpose Local Models - LFM 2.5: [lfm2.5-1.2b-instruct](https://huggingface.co/LiquidAI/LFM2.5-1.2B-Instruct-GGUF) - Ministral 3: [ministral-3-8b-instruct](https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512-GGUF) - Gemma 4: [gemma4-e2b-it](https://huggingface.co/unsloth/gemma-4-E2B-it-qat-GGUF), [gemma4-e4b-it](https://huggingface.co/unsloth/gemma-4-E4B-it-qat-GGUF), [gemma4-12b-it](https://huggingface.co/unsloth/gemma-4-12B-it-qat-GGUF), [gemma4-26b-a4b-it](https://huggingface.co/unsloth/gemma-4-26B-A4B-it-qat-GGUF), [gemma4-31b-it](https://huggingface.co/unsloth/gemma-4-31B-it-qat-GGUF) - Qwen 3.5: [qwen3.5-0.8b](https://huggingface.co/unsloth/Qwen3.5-0.8B-GGUF), [qwen3.5-2b](https://huggingface.co/unsloth/Qwen3.5-2B-GGUF), [qwen3.5-4b](https://huggingface.co/unsloth/Qwen3.5-4B-GGUF), [qwen3.5-9b](https://huggingface.co/unsloth/Qwen3.5-9B-GGUF), [qwen3.5-27b](https://huggingface.co/unsloth/Qwen3.5-27B-GGUF), [qwen3.5-35b-a3b](https://huggingface.co/unsloth/Qwen3.5-35B-A3B-GGUF) - Qwen 3.6: [qwen3.6-27b](https://huggingface.co/unsloth/Qwen3.6-27B-GGUF), [qwen3.6-35b-a3b](https://huggingface.co/unsloth/Qwen3.6-35B-A3B-GGUF) - Qwen 3.8: [qwen3.8-27b](https://huggingface.co/unsloth/Qwen3.8-27B-GGUF) #### Uncensored Local Models - Gemma 4 uncensored: [gemma4-e2b-uncensored](https://huggingface.co/HauhauCS/Gemma-4-E2B-Uncensored-HauhauCS-Aggressive), [gemma4-e4b-uncensored](https://huggingface.co/HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive), [gemma4-12b-uncensored](https://huggingface.co/HauhauCS/Gemma4-12B-QAT-Uncensored-HauhauCS-Balanced), [gemma4-26b-a4b-uncensored](https://huggingface.co/HauhauCS/Gemma4-26B-A4B-QAT-Uncensored-HauhauCS-Balanced-MTP), [gemma4-31b-uncensored](https://huggingface.co/HauhauCS/Gemma4-31B-QAT-Uncensored-HauhauCS-Balanced-MTP) - Qwen 3.5 uncensored: [qwen3.5-2b-uncensored](https://huggingface.co/HauhauCS/Qwen3.5-2B-Uncensored-HauhauCS-Aggressive), [qwen3.5-4b-uncensored](https://huggingface.co/HauhauCS/Qwen3.5-4B-Uncensored-HauhauCS-Aggressive), [qwen3.5-9b-uncensored](https://huggingface.co/HauhauCS/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive) - Qwen 3.6 uncensored: [qwen3.6-27b-uncensored](https://huggingface.co/HauhauCS/Qwen3.6-27B-Uncensored-HauhauCS-Balanced), [qwen3.6-35b-a3b-uncensored](https://huggingface.co/HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive) - Qwen 3.8 uncensored: [qwen3.8-27b-uncensored](https://huggingface.co/HauhauCS/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-MTP-GGUF) #### Cloud Providers Hosted LLM providers: [OpenAI](https://platform.openai.com/), [Gemini](https://ai.google.dev/), [Claude](https://www.anthropic.com/api), [Grok](https://docs.x.ai/developers), [MiniMax](https://platform.minimax.io/), [DeepSeek](https://platform.deepseek.com/), and [OpenRouter](https://openrouter.ai/). #### Machine Translation Providers Machine-translation providers: [DeepL](https://www.deepl.com/), [Google Cloud Translation](https://cloud.google.com/translate), and [Caiyun](https://fanyi.caiyunapp.com/). #### OpenAI-Compatible Providers OpenAI-compatible endpoints are also supported. ## Installation Download release builds from the [releases page](https://github.com/koharu-rs/koharu/releases/latest). [Installation requirements and first launch](https://koharu.rs/en/installation) vary by operating system. Builds are available for Windows, macOS, and Linux. ### WinGet Install on Windows with [winget](https://learn.microsoft.com/en-us/windows/package-manager/winget/): ```bash winget install koharu ``` ### Homebrew Install on macOS with [Homebrew](https://brew.sh/): ```bash brew install --cask koharu ``` ## Troubleshooting Startup, runtime, model, and provider errors are covered in [Troubleshooting](https://koharu.rs/en/reference/troubleshooting). Set `RUST_LOG` to `debug` or `trace` for verbose logs: ```bash # macOS / Linux RUST_LOG=debug koharu # Windows (PowerShell) $env:RUST_LOG="debug"; koharu.exe ``` ## Development Platform dependencies and validation commands for local builds are listed in [Development Setup](https://koharu.rs/en/development/setup). ### Prerequisites - [Rust](https://www.rust-lang.org/tools/install) 1.97.1 or later (Rust 2024 edition) - [Bun](https://bun.sh/) 1.3.14 or later - [LLVM](https://llvm.org/) 22.1.8 or later ### Install dependencies ```bash bun install ``` ### Development ```bash bun dev ``` ### Build ```bash bun run build ``` The executable is written to `target/release`. ## Sponsorship If Koharu is useful in your workflow, consider sponsoring the project. - [GitHub Sponsors](https://github.com/sponsors/mayocream) - [Patreon](https://www.patreon.com/mayocream) ## Contributors ❤️ Thanks to all the contributors who have helped make Koharu better! ## License Copyright 2025-2026 Mayo Takanashi and Koharu contributors. Koharu is dual-licensed under the [MIT License](LICENSE-MIT) or the [Apache License, Version 2.0](LICENSE-APACHE), at your option.