This session will provide hands-on training in the W3C Multimodal Architecture using interoperable components developed by Openstream, Loquendo, and other participants in the W3C Multimodal Interaction Working Group. The session will overview Openstream’s CueMe implementation of the W3C multimodal architecture and Loquendo’s MRCP speech server. Participants will be provided with a research version of the CueMe platform, documentation, and development tools that will be used in a hands-on multimodal development exercise. Both client-server and offline component distribution models will be discussed.