Ontology-Based Data Access has been studied so far for relational structures and deployed on top of relational databases. This paradigm enables a uniform access to heterogeneous data sources, also coping with incomplete information. Whether OBDA is suitable also for non-relational structures, like those shared by increasingly popular NOSQL languages, is still an open question. In this paper, we study the problem of answering ontology-mediated queries on top of key-value stores. We formalize the data model and core queries of these systems, and introduce a rule language to express lightweight ontologies on top of data. We study the decidability and data complexity of query answering in this setting.