To parse the neural basis of motor behavior, we leveraged the unique swim-to-walk transition of the Xenopus metamorphosis. This transition enabled us to dissect the spinal circuitry underlying tail and limb-based movement. We combined a robust behavioral assay with molecular cell-type profiling in wild-type, and then loss-of-function mutant, tadpoles to evaluate how neuronal diversity scaled with and contributed to the observed increase in movement complexity.We devised a high-speed behavioral setup to track frog’s body parts, quantify complex movement, and create locomotor profiles correlated with spinal cell-type composition. We profiled the expression of transcription factors that define motor neurons (MNs) and V1 interneurons (V1s), an inhibitory class modulating MN firing. In free-swimming tadpoles, both doubled in number and diversify in their transcriptional profile, acquiring subpopulations resembling the mouse hypaxial and preganglionic motor columns and V1 clades. At metamorphosis, spinal neuron diversity peaked with limb motor column formation and increased V1 number and diversity, matching the neonate mouse.Finally, we investigated how loss of a cell type altered the frog’s motor repertoire. Using CRISPR/Cas9, we selectively knocked out three master regulator genes: FoxP1, Engrailed-1 and Prdm12. By comparing the behavior of mutant versus wild-type animals, we unravel how each subtype contributes to tail and limb function. Our work maps MN and V1 molecular properties onto tail and limb-based behavior of frog metamorphosis, defining how transcriptional diversity scales with movement complexity and demonstrating its conservation across tetrapods.