If P is a lattice polytope (i.e., P is the convex hull of finitely many integer points in R d ), Ehrhart's famous theorem asserts that the integer-point counting function |nP X Z d | is a polynomial in the integer variable n. Equivalently, the generating function 1 `řně1 |nP X Z d | z n is a rational function of the form h ˚pzq p1´zq d`1 ; we call h ˚pzq the h ˚polynomial of P . There are several known necessary conditions for h ˚-polynomials, including results by Hibi, Stanley, and Stapledon, who used an interplay of arithmetic (integer-point structure) and topological (local h-vectors of triangulations) data of a given polytope. We introduce an alternative ansatz to understand Ehrhart theory through the h ˚-polynomial of the boundary of a polytope, recovering all of the above results and their extensions for rational polytopes in a unifying manner. We include applications for (rational) Gorenstein polytopes and rational Ehrhart dilations.