Particle dark matter could have a mass anywhere from that of ultralight candidates, m χ ∼ 10 −21 eV, to scales well above the GeV. Conventional laboratory searches are sensitive to a range of masses close to the weak scale, while new techniques are required to explore candidates outside this realm. In particular lighter candidates are difficult to detect due to their small momentum. Here we study two experimental set-ups which do not require transfer of momentum to detect dark matter: atomic clocks and co-magnetometers. These experiments probe dark matter that couples to the spin of matter via the very precise measurement of the energy difference between atomic states of different angular momenta. This coupling is possible (even natural) in most dark matter models, and we translate the current experimental sensitivity into implications for different dark matter models. It is found that the constraints from current atomic clocks and co-magnetometers can be competitive in the mass range m χ ∼ 10 −21 − 10 3 eV, depending on the model. We also comment on the (negligible) effect of different astrophysical neutrino backgrounds. 3 We take here the theory after EWSB, otherwise e L couplings come together with ν s and G u L = G d L . 4 For example the operatorψγ µ γ 5 ψ∂ µ χ 2 is proportional to the transferred momentum q so we discard it, but ∂ µ χ 2 is the only current that can be built with a real field; in contrast for a complex scalar we have two: ∂ µ (χ † χ) and (∂ µ χ † )χ − χ † ∂ µ χ. The second one is proportional to the sum of incoming and outgoing DM momenta in a scattering process.