A system and method is disclosed for exposing complete call flow speech components to users via Web services that are called by converged voice-Web server applications. The Web service call flow speech components may be accessed by application server pages resident on a Web server and accessible through either a telephony environment or a data network environment, such as the Internet. The Web service call flow speech objects automatically generate the appropriate speech access formatting into the corresponding server application page. By removing the speech component programming from development of the application server page, the representative embodiments expand the potential use of multimodal and or converged voice-Web applications and devices.