Our overall goal is to provide a full-featured integration of Asterisk and OpenVXI including as many features as possible while conforming to the VoiceXML standard.

Currently the feature most lacking is speech recognition. We are tackling this now. One approach is to make use of the Asterisk speech API (new in version 1.4) that currently is integrated with LumenVox speech recognition at the back end. This approach has the benefit that Lumenvox is already integrated. We’re tackling this approach first.

Another approach is to integrate sphinx2 with VoiceGlue. This has the advantage of keeping the entire product under the GPL license. Our goal is to perform this integration as well.

Further goals include more testing and support for the VoiceXML standard. It would be nice to be able to pass the VoiceXML conformace tests that are out there. Also, it would be fun to try to implement a lightweight version of this integration, all in C code, that could run on embedded devices.