We present the concept of a theory machine, which is an atemporal computational formalism that is deployable within an arbitrary logical system. Theory machines are intended to capture computation on an arbitrary system, both physical and unphysical, including quantum computers, Blum-Shub-Smale machines, and infinite time Turing machines. We demonstrate that for finite problems, the computational power of any device characterisable by a finite first-order theory machine is equivalent to that of a Turing machine. Whereas for infinite problems, their computational power is equivalent to that of a type-2 machine. We then develop a concept of complexity for theory machines, and prove that the class of problems decidable by a finite first order theory machine with polynomial resources is equal to 𝒩𝒫 ∩ co-𝒩𝒫.