Disfluency is verbally expressed by several markers (filled, unfilled pauses, repetitions, selfrepairs,etc). This study is grounded in the functionally ambivalent view of (Dis)fluencyfollowing Crible, (2017) and Götz (2013), but with a multimodal and interactional approach.Previous research has shown a coordination between speech and gesture suspension (Gullberg,2013, 2018; Seyfedinnipur 2006). The aim of our paper is thus to examine how (dis)fluentspeech and gestures can be synchronized, and how visual-gestural features can provide a finerunderstanding of (dis)fluency. Our analyses are conducted on 3 pairs of French and Americanspeakers interacting both in their L1 and their L2. (Dis)fluency markers were annotatedaccording to their multimodal features. Qualitative analyses revealed how the notions of timesuspension and planning associated with (dis)fluency were also found in gesture. This stronglysupports the idea that (dis)fluency is to be considered a multimodal phenomenon, and its visualcues are essential for a closer examination of its pragmatic functions.