Modeling stochastic programs involves manipulating several objects in addition to those in deterministic mathematical programs. Many extensions to Algebraic Modeling Languages (AML) have been proposed to deal with these objects, but none have been implemented. We argue that some of these objects do not lend themselves well to a textual representation. Thus, we propose a system which is an addition, as opposed an extension, to an AML. The system handles some objects outside the AML and integrates all the tools to create a solvable instance. In addition, the system covers all aspects of modeling, from problem definition and model elicitation down to generating different formulations to be used by different solution methods from the same model.