The properties of planet-forming midplanes of protostellar disks remain largely unprobed by observations due to the high optical depths of common molecular lines. However, rotational emission lines from rare isotopologues may have optical depth near unity in the vertical direction, so that the lines are strong enough to be detected, yet remain transparent enough to trace the disk midplane. We have computed chemical models of protostellar disks including different C and O isotopes and detailed photochemical reactions. The CO condensation front is in the giant planet-forming region, within 20 AU of the star. We show that the optical depths of low-order rotational lines of C17O are around unity, which suggests that it may be possible to see into the disk midplane using C17O. In lower-mass disks, the slightly more abundant C18O is a possible midplane probe. ALMA observations would provide estimates of the disk midplane temperature if CO ice line were spatially resolved. With our computed C17O/H2 abundance ratio, we would also be able to measure disk surface densities from the fluxes of low-order C17O transitions.