A *small semantic database* is a mathematical database where each entry is submitted by a human contributor, and where each entry includes supporting metadata allowing software to provide some level of automated deduction and/or search. In this talk we will explore the model used by the pi-Base community database of topological counterexamples. In this model, the database tracks a category of objects and their properties, with theorems that allow for automated deduction (e.g. given the theorem P ∧ Q ⇒ R, an object tagged with properties P and ¬R is automatically tagged with ¬Q).