• Excellent recognition with little training
  • Works in English, French, German, and Italian
  • Works in all applications
  • Users can train individual words, as well as correct errors in recognition


  • High learning-curve
  • Can use a lot of memory and processor cycles
  • Not all words can be effectively trained
  • No printable list of all commands means users must memorize many commands
  • Editing is not user-friendly, and is fraught with errors