Abstract
Katukinan, Arawan, and Harakmbut are small language families spoken in south-western Amazonia. These families have received some attention, but there are no consistently transcribed and machine-readable datasets available for them. We address this lacuna by introducing the first publicly available linguistic dataset of Arawan languages as the first part of the Katukinan-Arawan-Harakmbut Database, created with the goal of providing and regularly updating a list of lexical items in a consistent transcription and with cognacy annotation. The database is being developed to be used in quantitative and genealogical investigations.