Voltage-gated potassium (Kv) channels modulate the function of electrically-excitable and non-excitable cells by using several types of gates to regulate ion flow through the channels. An important gating mechanism, C-type inactivation, limits ion flow by transitioning Kv channels into a non-conducting inactivated state. Here, we highlight two recent papers, one on the human Kv1.3 channel and the second on the Drosophila Shaker Kv channel, that combined cryogenic electron microscopy and molecular dynamics simulation to define mechanisms underlying C-type inactivation. In both channels, the transition to the non-conducting inactivated conformation begins with the rupture of an intra-subunit hydrogen bond that fastens the selectivity filter to the pore helix. The freed filter swings outwards and gets tethered to an external residue. As a result, the extracellular end of the selectivity filter dilates and K+ permeation through the pore is impaired. Recovery from inactivation may entail a reversal of this process. Such a reversal, at least partially, is induced by the peptide dalazatide. Binding of dalazatide to external residues in Kv1.3 frees the filter to swing inwards. The extracellular end of the selectivity filter narrows allowing K+ to move in single file through the pore typical of conventional knock-on conduction. Inter-subunit hydrogen bonds that stabilize the outer pore in the dalazatide-bound structure are equivalent to those in open-conducting conformations of Kv channels. However, the intra-subunit bond that fastens the filter to the pore-helix is absent, suggesting an incomplete reversal of the process. These mechanisms define how Kv channels self-regulate the flow of K+ by changing the conformation of the selectivity filter.