This paper gives an overview of the Moses tool suite, a set of tools for visual language programming. In Moses, visual language syntax is defined by first-order predicates over the abstract syntax of a picture, represented by an attributed graph. One way of specifying language semantics in Moses is by writing an Abstract State Machine that interprets a given attributed graph. This paper shows how the editor is parameterized with a description of a visual language, and discusses briefly the generic architecture used to animate and debug visual programs.