In this paper we argue that, even though there are strong theoretical and empirical reasons to expect a violation of spatial isotropy at short distances, contemporary setups for probing gravitational interactions at short distances have not been configured to measure such spatial anisotropies. We propose a simple modification to the state-of-the-art torsion pendulum design and numerically demonstrate that it suppresses signals due to the large spatially-isotropic component of the gravitational force while maintaining a high sensitivity to short-range spatial anisotropies. We incorporate anisotropy using both Yukawa-type and power-law-type short-distance corrections to gravity. The proposed differential torsion pendulum is shown to be capable of making sensitive measurements of small gravitational anisotropies and the resulting anisotropic torques are largely independent of the details of the underlying short-distance modification to gravity. Thus, if there is an anisotropic modification to gravity, from any theory, in any form of the modified potential, the proposed setup provides a practical means of detecting it.