Our contribution compares various types of escape rooms (ERs) and argues for a theoretical Escape Room Format (ERF). We scrutinise analogue rooms, 2D computer games and virtual reality (VR) escape rooms by relying on Lars Elleström’s model of intermediality. Elleström distinguishes between four modalities of media: the material, the sensory, the spatiotemporal and the semiotic modalities, and the case studies discuss escape rooms in terms of how these modalities are combined for a coherent immersive experience. We point out the main impact the different media types have on articulating immersion: how physical escape rooms are more bounded by strict causal sequencing and the materiality of design, while VR escape rooms offer less predictable puzzles but a more direct way of outlining the narrative.