Hi everyone,

I've been doing a lot of research into all things VoIP lately:
asterisk, voicexml, the JAIN API, OpenVXI, etc., etc. Having gained a
basic appreciation of these various technologies, I was hoping someone
could offer some insight into which option would best suit my goal:

I need to build a relatively basic server architecture that can accept
SIP calls from a remote user, pass audio (from the user) to a
parser/recognizer/etc., and then returns pre-recorded audio and
language recognition "scores". The usual features such as barge-in and
silence-detection are preferred, but no sort of routing or PBX features
are needed.

The catch is, we are trying to do it with free tools, with a particular
emphasis on implementing the Sphinx recognizer. The description of what
I want to do is exactly what voicexml provides, but I have yet to find
a voicexml solution in which I can incorporate Sphinx (and it seems to
me that OpenVXI is a mess). I have looked into building a custom server
using the Galaxy architecture, but I am having trouble finding
information on how to establish a physical SIP connection. JAIN looks
great, but unless Im mistaken, more lower-level work must be done to
establish a connection (SIP stack?).

I would really appreciate it if someone with some experience would
offer a helping hand and give their impressions. Is it possible to
implement voicexml with Sphinx? Would working from the group up with an
SIP API be better? Could I strip down Asterisk to do the job?

Thanks a lot.