The goal of this work is to design a machine translation system for a low-resource family of dialects, collectively known as Swiss German. We list the parallel resources that we collected, and present three strategies for normalizing Swiss German input in order to address the regional