Installation

Getting the Manhwa OCR Tool installed is the first step. For the best stability and performance, we strongly recommend running the application directly from its source code.

1. From Source Code (Recommended Method)

This method ensures you have the latest stable version and the correct dependencies for your system. You will need Python (3.8+ recommended) and pip installed.

Step 1: Install PyTorch (Crucial First Step)

The application's OCR engine relies on PyTorch. It must be installed separately before other dependencies to ensure you get the version compatible with your hardware (especially if you have an NVIDIA GPU for CUDA acceleration).

  1. Go to the Official PyTorch Website: Open your browser and navigate to pytorch.org.
  2. Use the Install Helper: On the homepage, you will find a configuration tool. Select your preferences (e.g., Stable, your OS, Pip, Python, your CUDA version or CPU).
  3. Copy and Run the Command: The website will generate a specific pip install command for you. Copy it and run it in your terminal.

Example Command (Do NOT blindly copy this!) The command generated for you might look something like this for a system with CUDA 12.1:

pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu121

Your command will likely be different. Always use the command generated by the official PyTorch website.

Step 2: Clone the Repository

Once PyTorch is installed, open your terminal and clone the project's GitHub repository:

git clone https://github.com/kha-white/manhwa-ocr-tool.git
cd manhwa-ocr-tool

(Note: Replace the URL with the official repository URL if it differs.)

Step 3: Install Other Dependencies

Now, install the remaining required Python libraries using the requirements.txt file:

pip install -r requirements.txt

Step 4: Run the Application

Launch the tool from your terminal:

python main.py

2. Via Executable (Alpha / Unstable)

Warning: Pre-compiled executables are currently in an alpha stage. They are intended for testing only and may be broken or experience critical errors (like import errors on launch). For a stable experience, please use the "From Source Code" method above.

If you wish to test an executable:

  1. Download: Go to the official GitHub releases page and download the latest package for your operating system (Windows or macOS).
  2. Install:
    • Windows: Run the .exe installer or extract the portable .zip file.
    • macOS: Open the .dmg file and drag the icon into your Applications folder.

Common Installation Issues

GUI Library Conflict (PyQt5 vs. PySide6)

The application is built with PySide6. If you have the conflicting PyQt5 library installed, it will not start. The tool is designed to detect this and will show a detailed error message with a "Copy Commands" button.

If you encounter this, follow the on-screen instructions or see our Troubleshooting Guide for a step-by-step solution.