Despite the long history of dark matter, its nature is still unknown. Cold dark matter (CDM) remains the generally accepted working hypothesis. Apparent shortcomings in the ability of CDM to explain various observations have been noted and have led to the development of alternative hypotheses, but so far none have been able to dethrone CDM. This is partly due to the presence of baryons in galaxies, whose feedback processes and radiative properties are far more complex than the physics of dark matter. In the presence of baryonic feedback, many dark-matter models start losing their distinctive profiles, leaving us unable to distinguish between them. One promising way out of this conundrum is to study dark matter in environments with as few baryons as possible. Ultra-faint dwarf galaxies (UFDs) are the faintest, least massive, and most dark matter–dominated galaxies known. They are predicted to have dark-matter distributions unchanged by baryonic feedback. In this colloquium I will present the current state of research to constrain the nature and properties of dark matter using UFDs, including the first results from a novel 100-hour MUSE survey of UFDs. I will address the constraints on primordial black holes from their dynamical effects on stellar distributions, as well as the constraints on various types of particulate dark matter (weakly interacting massive particles, axion-like particles, self-interacting dark matter, and fuzzy dark matter) from emission-line searches and the first determined dark matter–density profile of a UFD. I will end with an outlook for the near future of this field.