You say that your questions are already in Voicexml format. Does that mean
you already have <field> tags containing the text to speak and the allowable
caller entries? If so, then the easiest thing may be to use the
subdialog_invoke element to invoke your vxml as a subdialog and let the vxml
it creates go directly to the gateway.
Otherwise, you will have to write some code to grab the tts-text for the
prompt from the vxml returned by your url, store it into session data var
and then use that.
As geoff mentions, if the prompts can be pre-recorded, you can have the url
return the prompt name into a session var and use that in the audio URI
field.
In any case, you have 3 things to do.
A) write an element (action or decision is easiest) to exec the url and grab
the tts or file name and store it into a variable.

use that. Variable in the audio tab of some element that collects input
from the caller (i would use a form element for flexibility), either in the
audio file or tts box.
C)decide what the allowable entries from the caller are and indicate that
either in the settings tab, or after youve collected the input.
Janine
Sent from my Verizon Wireless Phone
-----Original message-----
From: Cisco Developer Community Forums <cdicuser@developer.cisco.com>
To: "cdicuser@developer.cisco.com" <cdicuser@developer.cisco.com>
Sent: Fri, Mar 9, 2012 17:37:52 GMT+00:00
Subject: New Message from Phani Venigalla in Customer Voice Portal (CVP) -
General Discussion - All Versions: RE: classpath jar problem with Spring
Framework on CVP application