China’s southwestern province of Sichuan is a linguistic terra incognita. It hosts a multitude of diverse languages, most of which are at best rudimentarily documented and described. The four Tibeto-Burman languages investigated by our project—Duoxu, Lizu, Ersu and Xumi—encapsulate the hallmarks of the area: diversity, complexity, a lack of systematic data and research, and a high level of language endangerment. While all four languages are endangered, the urgency of documentation is most acute in the case of Duoxu. This language is spoken by no more than a handful of members of the oldest generation, and it is no longer transmitted to children. Adding to the sense of urgency is that Duoxu is one of the least-studied Tibeto-Burman languages of China. Although this language has been repeatedly surveyed in the past sixty years, very little data from those surveys has been released. As a result, even basic information on Duoxu was previously lacking, such as the total number of remaining speakers, their level of proficiency, and the domains in which the language is used.