These are the document models used for MModal voice recognition. These document models are configured in AnyModal Console and define the rules for VR drafts, i.e., the number of spaces after a period, etc. The appropriate document model is configured in the Form setup under the Voice Rec tab.