The purpose of this paper is to investigate temporal variations in the northward, X, eastward, Y, and downward, Z, components of the geomagnetic field recorded during the October 14, 2023 annular solar eclipse, which main features include its annularity, the eclipse occurrence from local dawn to local dusk, its magnitude variation from 0.30 to 0.86, and the longest ever-observed path across the mainland of the Americas, covering latitudes from ∼65°N to 12°S. The analysis was made possible due to the data on temporal variations in the northward, X, eastward, Y, and downward, Z, components of the geomagnetic field collected at thirteen International Real-time Magnetic Observatory Network magnetometer stations (https://imag-data.bgs.ac.uk/GIN_V1/GINForms2). The solar eclipse acted to cause non-sinusoidal and quasi-sinusoidal perturbations having temporal durations of 180–240 min in all geomagnetic field components on a global scale (∼8000 km). The X-component experienced the largest perturbations attaining 10–20 nT, and the Z-component underwent the smallest disturbances. The quasi-sinusoidal perturbation amplitude did not exceed 5–6 nT, and the period most often showed variations within 15–40 min. The magnetic effect exhibited a tendency to increase with solar eclipse magnitude, while the magnitude of the effect has been shown to be significantly dependent on geographic coordinates, local time, ionospheric state, and the patterns of ionospheric currents as well. During the solar eclipse, the electron density depletion was estimated to be ∼0.10 to ∼0.40–0.60 when the eclipse obscuration Amax varied from 19% to 82%. The movement of the lunar shadow was accompanied by the generation of atmospheric gravity waves with period of ∼10–80 min and by electron density perturbations with amplitudes of the order of 0.01–0.03. The estimates made on the assumption that the magnetic effect is due to the ionospheric current disruptions show good agreement with the observations.