YomiNinja  by matt-m-o

Open-source OCR tool for language learners

created 1 year ago
474 stars

Top 65.3% on sourcepulse

GitHubView on GitHub
Project Summary

YomiNinja is an open-source OCR and dictionary tool designed for language learners, enabling efficient text extraction from visual content like games and videos. It overlays extracted text onto the original media, facilitating quick look-ups with integrated pop-up dictionaries, thereby minimizing distractions and streamlining vocabulary acquisition.

How It Works

YomiNinja leverages Optical Character Recognition (OCR) to capture text from the screen or specific windows. It then integrates with browser-based dictionary extensions (like 10Ten and Yomitan) to provide instant definitions for unfamiliar words. The tool supports OCR templates for optimizing text extraction and includes features like auto-OCR and text-to-speech.

Quick Start & Requirements

  • Installation: Download and install the latest release for Windows (.exe), Linux (.deb, .rpm), or macOS (.dmg).
  • Prerequisites:
    • Windows: Windows 10/11, VCRedist. Media Feature Pack required for N/KN editions.
    • Linux: X11 window system, xdotool. Wayland not supported.
    • macOS: Apple Vision OCR engine (language support depends on macOS version).
  • Build: Requires cloning the repository, building OCR services (Python gRPC), downloading PPOCR-Inference-Service, installing Node.js modules (npm install --force), and generating gRPC types (npm run grpc-types).
  • Resources: Building involves several steps and external dependencies.
  • Links: Demonstrations: 10ten, Yomichan.

Highlighted Details

  • Supports text extraction from the entire screen or specific windows.
  • Features built-in pop-up dictionaries and partial Chrome Extension support.
  • Includes OCR templates for optimizing efficiency and auto-OCR functionality.
  • Offers text-to-speech and WebSocket for Texthookers.

Maintenance & Community

  • Planned features include Anki integration, history, text translation, and support for more OCR/TTS engines.
  • Inspired by Yomichan, mokuro, and kanjitomo-ocr.

Licensing & Compatibility

  • License: Not explicitly stated in the README.
  • Compatibility: Primarily targets Windows, Linux (X11), and macOS. ARM64 support is planned.

Limitations & Caveats

  • Wayland on Linux is not supported due to limitations with global shortcuts and window positioning.
  • Native ARM64 support for macOS is planned for version 0.8 and above.
  • Chrome Extension support is partial; only 10Ten and Yomitan are confirmed as pre-installed options.
Health Check
Last commit

1 day ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
1
Star History
46 stars in the last 90 days

Explore Similar Projects

Feedback? Help us improve.