Semiphysical modelling is often interpreted as an application of system identification where physical insight into the application is used to come up with suitable non-linear transformations of the raw measurements so as to allow for a good model structure. This modelling procedure is less ‘ambitious’ than those used for traditional physical modelling in that no complete physical structure is sought, just suitable inputs and outputs that can be subjected to more or less standard model structures such as linear regressions. In this paper we discuss a semiphysical modelling procedure and various tools supporting it. These include constructive algorithms originating from commutative and differential algebra as well as more informal tools such as the programming environment.