The "torus" is the central element of the most popular theory unifying various classes of AGNs, but it is usually described as "putative" because it has not been imaged yet. Since it is too small to be resolved with single-dish telescopes, one can only make indirect assumptions about its structure using models. Using infrared interferometry, however, we were able to resolve the circum-nuclear dust distributions for several nearby AGNs and achieved constraints on some further two dozen sources. We discovered circum-nuclear dust on parsec scales in all sources and, in two nearby sources, were able to dissect this dust into two distinct components. The compact component, a very thin disk, appears to be connected to the maser disk and the extended one, which is responsible for most of the mid-IR flux, is oriented perpendicularly to the circum-nuclear gas disks. What may come as a surprise when having in mind the standard unification cartoon actually connects well to observations on larger scales. Optically thin dust in the polar region, perhaps driven by a disk wind, could solve both the scale height problem of the torus and explain the missing anisotropy in the mid-IR − X-ray relation.