In this study, we discuss producing radioisotopes using linear electron accelerators and address production and separation issues of photoneutron (γ,n) and photoproton (γ,p) reactions. While (γ,n) reactions typically result in greater yields, separating product nuclides from the target is challenging since the chemical properties of both are the same. Yields of (γ,p) reactions are typically lower than (γ,n) ones, however they have the advantage that target and product nuclides belong to different chemical species so their separation is often not such an intricate problem. In this paper we consider two examples, (100)Mo(γ,n)(99)Mo and (68)Zn(γ,p)(67)Cu, of photonuclear reactions. Monte-Carlo simulations of the yields are benchmarked with experimental data obtained at the Idaho Accelerator Center using a 44MeV linear electron accelerator. We propose using a kinematic recoil method for photoneutron production. This technique requires (100)Mo target material to be in the form of nanoparticles coated with a catcher material. During irradiation, (99)Mo atoms recoil and get trapped in the coating layer. After irradiation, the coating is dissolved and (99)Mo is collected. At the same time, (100)Mo nanoparticles can be reused. For the photoproduction method, (67)Cu can be separated from the target nuclides, (68)Zn, using standard exchange chromatography methods. Monte-Carlo simulations were performed and the (99)Mo activity was predicted to be about 7MBq/(g(⁎)kW(⁎)h) while (67)Cu activity was predicted to be about 1MBq/(g(⁎)kW(⁎)h). Experimental data confirm the predicted activity for both cases which proves that photonuclear reactions can be used to produce radioisotopes. Lists of medical isotopes which might be obtained using photonuclear reactions have been compiled and are included as well.