The eciency of thermoelectric devices has not increased signi®cantly during the last 40 years, and currently, the ®gure-of-merit (ZT) of the best materials is less than one. However, there has been a resurgence of interest in thermoelectric materials due to recent results that have shown that size-eects in nanostructured materials may enable much more ecient thermoelectric devices. Theoretical calculations for isolated single nanowires predict over a 10-fold increase in the ®gure-of-merit. In order to realize these size-eects, wires must be synthesized with diameters less than the thermal de Broglie wavelength (typically less than 10 nm) which is smaller than is currently accessible by lithographic techniques. Also, in order to transfer meaningful amounts of thermal energy, a practical device must consist of an array of a large number of nanowires in parallel. One way to potentially synthesize these nanowire arrays is to use the pores of microporous and mesoporous materials as a mold to template the diameter and orientation of wires of suitable thermoelectric materials. Here we provide a review of the relevant physics and derive a model to assess the feasibility of using microporous and mesoporous frameworks in the fabrication of thermoelectric devices. The model accounts for the possible deleterious eects of thermal conduction through the microporous or mesoporous framework and the presence of bulk thermoelectric material (that may result from defects in the microporous or mesoporous framework) in parallel with the wires on the thermoelectric ®gure-of-merit. Simulation results are reported for SBA-15, MCM-41, and zeolites VFI, LTL, and LTA embedded with Bi 2 Te 3 nanowires. The results show that the microporous frameworks may yield ®gures-of-merit much larger than one while thermal conduction through most mesoporous frameworks reduces the ®gure-of-merit below the level of currently available using traditional thermoelectric materials.