The performance of an architectural object is highly difficult to both define and measure in its complexity since it is integrating a constantly increasing amount of information, from concrete measurable characteristics to the subjective perception of individual users. The question arising though is how to predict the performance of a building and influence the design in order to increase it according to a significantly high number of criteria.The presented paper proposes two design tools, both developed and programmed in rhino python for the generation of freeform geometries. The tools are generated for specific tasks, but may be interpreted as exemplary for a way of defining and structuring a design program in order to increase its efficiency. Both tools rely on a computational core that is generally defined and may be fed with as many and different constraints and criteria as considered suitable for the defined task.