We propose a new point of view on multidimensional continued fraction algorithms inspired by Rauzy induction. The generic behaviour of such an algorithm is described here as a random walk on a graph that we call simplicial system. These systems provide a family of examples for random walks with memory recorded by a finite dimensional vector. We introduce a general criterion on these graphs that induces ergodicity together with a bundle of many other dynamical properties. In particular, after computing the representation of Brun, Selmer and Arnoux-Rauzy-Poincare algorithm in this formalism, it provides a unified proof of ergodicity for these classical examples as well as new results such as uniqueness of the measure of maximal entropy on a canonical suspension. These objects also bring a new perspective to some fractal sets such as Rauzy gaskets. We show general explicit upper bound on Hausdorff dimensions of fractals described in this formalism as well as a construction of their measure of maximal entropy. This implies in particular that the Rauzy gasket in all dimensions has Hausdorff dimension strictly smaller than its ambient space, as well as sharper bounds on the dimension and an asymptotic result.