Hira accepts, as argument text or as standard input, romaji text. It copies the text to standard output, converting the text that matches the romaji format to hiragana. Note that the output may not be accurate because of false matches and the inability to generate things like the tsu consonant-extending symbol.