The authors present a system that computes for a query that may be incomplete, complete approximations from above and from below. They assume a setting where queries are posed over a partially complete database, that is, a database that is generally incomplete, but is known to contain complete information about specific aspects of its application domain. Which parts are complete, is described by a set of so-called table-completeness statements. Previous paper led to a theoretical framework and an implementation that allowed one to determine whether in such a scenario a given conjunctive query is guaranteed to return a complete set of answers or not.