Database queries involving imprecise or fuzzy categories are an evolving area of academic research nowadays. In this paper we propose a novel approach to process queries purely expressed in natural language and demonstrate the effectiveness of using the concepts of Computing with Words in it. Currently, the system can process natural queries of single English sentences and retrieve the corresponding records. What is more, it allows to rank the output tuples by their importance. Such type of interface is much more natural for users, since they are free of using exotic crisp phrases that all standard query mechanisms provide.