This paper presents the first steps in the process of creating a multimedia corpus for the severely endangered Finland-Swedish Sign Language (FinSSL). In the paper, we will first outline the history and current situation of FinSSL and then move on to describe some of the foundational choices which we have made both in the earlier data collection and at the start of the currently ongoing annotation work. Finally, we will bring up challenges related to the corpus data processing and discuss the future uses of the corpus, especially from the point of view of the FinSSL revitalization process.