Abstract Through the simultaneous presence of graphically composed and materially existing elements, augmented reality (AR) offers ephemeral digital content that is the result of the momentary and, thus, unrepeatable alignment of a physical body and world and an AR system. Capturing the performative and embodied angles of screen-based AR through a combined film-analytical and cognitive lens, this paper focuses on how interfaces, content, and AR-manipulated bodies serve as apparatus for cinematic composition as well as storytelling and user engagement. Observing interactions with AR filters and backgrounds, we reflect on how users’ bodies and expressions that are mirrored on screen are translated into an immersive digital storyworld that exists in the temporal and spatial context of the AR experience and the related technology. AR filters and backgrounds’ affective quality, thus, lies in bodily control and in the creative act of choosing and moderating body characteristics, postures, and positions in real time in relation to the surrounding digitally manipulated or recorded environment. By moderating the representations of bodies and spaces as well as their interplay, AR users actively shape the visual composition of the on-screen space and, thereby, the visual narrative.