If you just want to synthesize japanese speech from japanese text, [nimopenjtalk](https://github.com/demotomohiro/nimopenjtalk) can do that. It is [Open JTalk](http://open-jtalk.sourceforge.net/) and [hts_engine API](http://hts-engine.sourceforge.net/) bindings for Nim. You can get wav file or wave data by calling Nim procedures with UTF8 string.
This example code just synthesizes speech from a Nim string: <https://github.com/demotomohiro/nimopenjtalk/blob/master/examples/speechSynthWave.nim> This example code take text from stdin and play the synthesized speech using openAL: <https://github.com/demotomohiro/nimopenjtalk/blob/master/examples/speechSynthOpenAL.nim>
