This paper contributes to the growing line of research that takes a multimodal approach in the study of interpreter-mediated dialogues. Drawing on insights from Conversation Analysis and multimodal analysis, it investigates how extended multi-unit turns unfold with interventions of an interpreter and, more specifically, what is the role of gaze in this process. The analysis is based on videos of interpreter-mediated dialogues (Dutch-Russian) recorded with mobile eye-tracking glasses. The paper argues that the interpreter’s gaze direction contributes both to the local management of turn-taking (next-speaker selection) and to sequence organization. More specifically, it shows how interpreter’s gaze orientation bears on the negotiation of possible transition relevance places and how it contributes to the smooth continuation of the projected extended multi-unit turn.