Extending AutoVoice with pocketsphinx

Are there any plans on extending AutoVoice to have its own voice recognition? I don't really like the Google offline recognition and always sending stuff to their server is not that great either.
I have looked for alternatives, apparently pocketsphinx is great, but noone seems to use it but RSen for his commandr and this app is just to cluttered for me. All i want is an app that takes what i say and pass it on to tasker.
Continueous listeing would of course be great but i mainly use conditions were i am prompted to talk anyways. So have you ever looked in that direction? Would you consider it? Or is it completely out of the question?
Thanks anyway for your great work =)

 

It has a seperate app called open mic+ which uses pocketsphinx instead of google voice for voice recognition and then triggers commandr

 

I only had it installed shortly but it seemed very smooth.. and from the comments of others it is also faster. Of course pocketsphinx needs a dictionary. I thought maybe it is possible to built an app in a way that you can change the dictionary and start of with a rather simple one. It will probably take a bit of time, but there will probably be tutorials how to build a dictionary and there will be a lot of custom ones, which means real customisation and also custom hot phrases^^
There are also some on voxforge, which can probably be used. And there already are some tutorials, i read over that topic during the christmas holidays and wanted to see if i could get it to work, but my java sucks and there is something wrong with the demo, that is not fixable when you have no idea what you are doing^^
http://cmusphinx.sourceforge.net/wiki/tutorialdict
http://cmusphinx.sourceforge.net/wiki/tutorialadapt