MMDAgent toolkit

Last update : August 9, 2013



MMDAgent is a toolkit for building voice interaction systems. The toolkit is released for contributing to the popularization of speech technology. Users can design users own dialog scenario, 3D agents, and voices. This software is released under the New and Simplified BSD license. Version 1.3.1 was released on December 25, 2012, and is available at the SourceForge website. The toolkit was created by the Department of Computer Science at the Nagoya Institute of Technology, Japan. The current members of the project team are Keiichi Tokuda, Akinobu Lee and Keiichiro Oura.

MMDAgent  employs MikuMikuDance (MDD) as a foundation for its 3D rendering system, as well as allowing users to maintain a lively conversation with their 3D companion. Hatsune Miku is one of the many models that can be used to hold a conversation with.

MMDAgent speech recognition

The speech recognition module of MMDAgent is based on Julius, the open-source large-vocabulary continuous speech recognition engine. New words to be recognized to MMDAgent can be added by making a user dictionary.

An MMDAgent WordPress Blog with news about the project was launched in August 2012.