Recent studies based on photometry from the Transiting Exoplanet Survey Satellite (TESS) have suggested that the dayside of KELT-1b, a strongly irradiated brown dwarf, is significantly brighter in visible light than what would be expected based on Spitzer observations in the infrared. We observed eight eclipses of KELT-1b with CHaracterising ExOPlanet Satellite (CHEOPS) to measure its dayside brightness temperature in the bluest passband observed so far, and we jointly modelled the CHEOPS photometry with the existing optical and near-infrared photometry from TESS, LBT, CFHT, and Spitzer. Our modelling has led to a self-consistent dayside spectrum for KELT-1b covering the CHEOPS, TESS, H, Ks, and Spitzer IRAC 3.6 and 4.5 µm bands, where our TESS, H, Ks, and Spitzer band estimates largely agree with the previous studies. However, we discovered a strong discrepancy between the CHEOPS and TESS bands. The CHEOPS observations yield a higher photometric precision than the TESS observations, but they do not show a significant eclipse signal, while a deep eclipse is detected in the TESS band. The derived TESS geometric albedo of 0.36−0.13+0.12 is difficult to reconcile with a CHEOPS geometric albedo that is consistent with zero because the two passbands have considerable overlap. Variability in cloud cover caused by the transport of transient nightside clouds to the dayside could provide an explanation for reconciling the TESS and CHEOPS geometric albedos, but this hypothesis needs to be tested by future observations.