Precise thermal neutron capture $\ensuremath{\gamma}$-ray cross sections ${\ensuremath{\sigma}}_{\ensuremath{\gamma}}$ for ${}^{39,40,41}$K were measured on a natural potassium target with the guided neutron beam at the Budapest Reactor. The cross sections were internally standardized using a stoichiometric KCl target with well-known ${}^{35}$Cl(n,$\ensuremath{\gamma}$) $\ensuremath{\gamma}$-ray cross sections [R\'evay and Moln\'ar, Radiochimica Acta 91, 361 (2003); Moln\'ar, R\'evay, and Belgya, Nucl. Instrum. Meth. Phys. Res. B 213, 32 (2004)]. These data were combined with $\ensuremath{\gamma}$-ray intensities from von Egidy et al. [von Egidy, Daniel, Hungerford, Schmidt, Lieb, Krusche, Kerr, Barreau, Borner, Brissot et al., J. Phys. G. Nucl. Phys. 10, 221 (1984)] and Krusche et al. [Krusche, Lieb, Ziegler, Daniel, von Egidy, Rascher, Barreau, Borner, and Warner, Nucl. Phys. A 417, 231 (1984); Krusche, Winter, Lieb, Hungerford, Schmidt, von Egidy, Scheerer, Kerr, and Borner, Nucl. Phys. A 439, 219 (1985)] to generate nearly complete capture $\ensuremath{\gamma}$-ray level schemes. Total radiative neutron cross sections were deduced from the total $\ensuremath{\gamma}$-ray cross section feeding the ground state, ${\ensuremath{\sigma}}_{0}=\ensuremath{\Sigma}{\ensuremath{\sigma}}_{\ensuremath{\gamma}}(\mathrm{GS})$ after correction for unobserved statistical $\ensuremath{\gamma}$-ray feeding from levels near the neutron capture energy. The corrections were performed with Monte Carlo simulations of the potassium thermal neutron capture decay schemes using the computer code dicebox where the simulated populations of low-lying levels are normalized to the measured cross section depopulating those levels. Comparisons of the simulated and experimental level feeding intensities have led to proposed new spins and parities for selected levels in the potassium isotopes where direct reactions are not a significant contribution. We determined the total radiative neutron cross sections ${\ensuremath{\sigma}}_{0}{(}^{39}\text{K})=2.28\ifmmode\pm\else\textpm\fi{}0.04$ b, ${\ensuremath{\sigma}}_{0}{(}^{40}\text{K})=90\ifmmode\pm\else\textpm\fi{}7$ b, and ${\ensuremath{\sigma}}_{0}{(}^{41}\text{K})=1.62\ifmmode\pm\else\textpm\fi{}0.03$ b from the prompt $\ensuremath{\gamma}$-ray data and the $\ensuremath{\gamma}$-ray transition probability ${P}_{\ensuremath{\gamma}}(1524.66)=0.164(4)$ in the ${\ensuremath{\beta}}^{\ensuremath{-}}$ decay of ${}^{42}$K in a low-background counting experiment.