The occipitotemporal cortex (OTC) contains areas that preferentially respond to specific object categories like body parts and tools, with a mirrored organization across its lateral and ventral parts. However, the exact role of these areas in supporting action or object recognition is still debated. Here, we investigate the selectivity and the multivariate representations of these selective clusters. We analysed an fMRI dataset in which participants were presented with images characterized by a different degree of action properties of either body parts (hands, whole-bodies, and faces) or objects (tools, manipulable objects, and non-manipulable objects). To test selectivity in OTC, we adopted a vector-of-ROI approach (Konkle & Caramazza, 2013), in which we created and extracted the activations from a series of partially overlapping spheres covering lateral and ventral OTC. Left lateral OTC exhibits a dorsal-posterior to ventral-anterior topographically organized action gradient, with selective and partly overlapping activations for bodies, hands, tools, and manipulable objects, and a selective overlap between hands and tools; left ventral OTC contains activations for hands distinct from the activations to faces and bodies. To further characterize this action gradient, we computed an action index capturing specific action-related components in OTC representational space. This revealed that the action-related activation gradient was reflected at the level of multivariate representations only in hand-selective clusters in both left lateral OTC and ventral OTC. Altogether, our results show that the object space in hand-selective visual cortex reflects action-related behaviourally-relevant dimensions, and suggests future paths to generate a computational model of this object space.