Julius は音声認識システムの開発・研究のためのオープンソースの高性能な汎用大語彙連続音声認識エンジンです。数万語彙の連続音声認識を一般のPC上でほぼ実時間で実行でき，認識率は90%以上です。言語モデル・音響モデルを任意に入れ替えられ，高い汎用性を持ちます．
[free version: does not contain grammar-based engine Julian]
(3.3p1)
Three bugs relating language model is fixed.
(3.3p2)
This release has three bug fixes and one new feature for
recording of recognized speech. See changelog for details.
Changes in 3.3p1-3.3p2:
- add new feature: "-record dir" records speech inputs sucessively into the specified directory.
- fix segfault on Solaris with "-input mfcfile".
- fix adin-cut bug when using module mode and adinnet together.
- fix output flush after last recognition output.
Changes in 3.3-3.3p1:
- (BIG!) fix misconfigure of default language model weights.
- fix output of wrong value in "-separatescore".
- fix dictionary reading (allow double space)
