polm/cutlet

Japanese to romaji converter in Python

What it solves

Cutlet provides a way to convert Japanese text into romaji (Latin alphabet) using various romanization systems, ensuring that proper nouns are capitalized and specific word exceptions are handled correctly.

How it works

It uses the fugashi library for tokenization and supports multiple romanization standards, including Modified Hepburn, Kunreisiki, and Nihonsiki. It can leverage UniDic dictionaries to use foreign spellings when available and allows users to define custom overrides for individual mappings or specific words.

Who it’s for

Developers who need to programmatically convert Japanese text to romaji for applications, URL slug generation, or linguistic analysis.

Highlights

  • Supports Modified Hepburn, Kunreisiki, and Nihonsiki systems.
  • Includes a built-in exceptions list for common names like Tokyo and Osaka.
  • Features a "slug mode" specifically for generating URL-friendly strings.
  • Allows custom overrides for specific words and mappings.
  • Integrates with UniDic for foreign spelling support.

Related

  • Project
    omdsh-dev/dsh-mnemon

    A composable memory system for DeepSeek Harness that organizes AI agent memory into runtime context, searchable documents, and long-term evidence.

  • Project
    KKKKhazix/sun-style-writing

    A writing skill for AI agents that implements ten restrained narrative techniques to replace sentimental AI prose with concrete, action-oriented storytelling.

  • Project
    GetStream/gemini-android

    A demonstration project showing how to integrate Google's Gemini AI with the Stream Chat SDK to build an AI-powered chat application on Android.

  • Project
    drfccv/mcp-server-12306

    An MCP server that enables AI assistants to query real-time ticket availability, pricing, and route information from the official 12306 Chinese railway system.