Abstract: A method and system for creating, using, and updating individual user voice portals in a multi-user environment is disclosed. Each user selects data sources, keywords and time intervals for searching the data sources, and grammar for accessing search results. Based on this information, an individual voice portal is created for the user. The requested information is extracted from the specified data sources and cached or stored in a database to increase extraction speed. The user accesses the individual voice portal using a unique identifier assigned to the user and the selected grammar. Because each user has a unique voice portal, the search space for software that interprets the grammar is decreased. As a result, the likelihood of misinterpretation and the time for extracting the requested information are decreased.