We propose a new universal framework for multi-stage decision making under limited information availability. It is developed as part of a larger research project which aims at providing analytical methods to compare and evaluate different models and algorithms for multi-stage decision making. In our setting, we have an open time horizon and limited information about upcoming stages in different granularity (e.g., deterministic information about the near future and/or forecast-based information about the more distant future). The framework is designed to reflect the need of decoupling uncertainty models from algorithms and to integrate standard disciplines, such as online optimization, stochastic programming and robust optimization, within a unified setting. Hence, the design of the framework provides the modeler with full flexibility to configure data availability, uncertainty representation, and solution methodology according to the problem under consideration. To the best of the authors’ knowl- edge, such a setting was not considered so far. The modeling capabilities of the framework are illustrated through a broad range of examples. In order to assess the quality of solution algorithms, the framework is connected to an evaluation module. The overall architecture consisting of the modeling framework, solution methodology, and evaluation module represents a comprehensive tool for comparing and relating different algorithmic strategies and different uncertainty models in multi-stage decision making problems.