All natural languages include visible expressions that communicate, including, for example, gesture, body posture and eye gaze. While the importance of documenting African languages is well established, few studies address the function, structure, and cultural significance of bodily actions in natural interactions. Obtaining video recordings of natural language use, especially by a non-local researcher, involves location specific considerations that are often left out in field guides written for linguistic researchers in general. In this paper we offer a few preparatory procedures for carrying out video recordings of spontaneous interactions in rural Malawi. We address logistical procedures before entering the research site, methods for recording natural interactions under the subjects' informed consent, considerations for selecting equipment and suggestions on data analysis. We argue that following local conventions for obtaining research permission, respecting and socio-cultural norms and participating in everyday community activities are essential for collecting ethically responsible and analytically useful data. While the paper focuses specifically on Malawi, its emphasis on practicing and integrating local social expectations is essential for carrying out video data collection in other African societies.