Methods for selecting acoustic model for use in a voice command platform

A voice command platform provides an application developer with the ability to select a particular acoustic model from a plurality of available acoustic models. In a representative embodiment, the application executing on the voice command platform is written in Voice XML, a tag-based language. The programmable selection of the speech recognition engine and the acoustic model is provided by means of a VXML root document having VXML code. The of the element of the VXML specification is extended to encompass selection of a particular acoustic model. The root document includes the element with the name of the acoustic model to use. Other methods for selection of acoustic model are also disclosed, including HTTP headers, interactive selection, and the use of area codes or local exchanges where the user using the voice application is calling from.