Physicists use term “ratchet” to describe systems which, at first glance, appear symmetrical but generate non-zero currents. In light of recent interest in such systems within the context of self-propelling (“active”) matter, we investigate a semi-Markov stochastic ratchet model that does not need external forces or potentials. Here, current is induced due to asymmetry in the shape of waiting-time distributions (i.e., memory mechanism in semi-Markov processes). The model considered is related to the popular stochastic reset archetype and is also reminiscent of the run-and-tumble framework. We show exact expression for mean ratchet current, analyse corresponding fluctuation behaviour and explore the possibility of dynamical phase transitions. To this end, we apply Laplace transforms, renewal theory and theory of large deviations.