Multiplatform Speech recognition

Hi list,
I need to develop a simple speech recognition app in max/msp. The app should just allow to be trained on a 30-40 words vocabulary and show the appropriate word. The important thing is training bacause it needs to be accessed and trained by different persons.
I would develop it on WIndows and then install it on Mac. I need to know if there is some good multiplatform object or some standard way of doing this (examples...), since I see there is no windows port for the listen~ object. Thank you,
Ale