We develop a novel statistical strong-lensing approach to probe the cosmological parameters by exploiting multiple redshift image systems behind galaxies or galaxy clusters. The method relies on free-form mass inversion of strong lenses and does not need any additional information other than gravitational lensing. Since in free-form lensing the solution space is a high-dimensional convex polytope, we consider Bayesian model comparison analysis to infer the cosmological parameters. The volume of the solution space is taken as a tracer of the probability of the underlying cosmological assumption. In contrast to parametric mass inversions, our method accounts for the mass-sheet degeneracy, which implies a degeneracy between the steepness of the profile and the cosmological parameters. Parametric models typically break this degeneracy, introducing hidden priors to the analysis that contaminate the inference of the parameters. We test our method with synthetic lenses, showing that it is able to infer the assumed cosmological parameters. Applied to the Cluster Lensing And Supernova survey with Hubble (CLASH) clusters, the method might be competitive with other probes.