Kana accepts, as argument text or as standard input, Japanese UTF-8 text. It copies the text to standard output, converting katakana and hiragana to romaji, leaving the rest (including kanji) unmodified.