Abstract Infragravity (IG) waves (periods nominally 25–250 s) are generated in shallow water by nonlinear difference‐frequency interactions between shorter period (nominally 4–25 s) incident sea‐swell waves. Previous studies show strong IG reflection from the beach face and detectable IG energy in alongshore‐propagating edge waves. Here, a linear wave model is developed to invert observed pressure (P) and velocity (cross‐shore and alongshore, U, V) cross‐spectra (including sensors separated in both the alongshore and cross‐shore directions) into cross‐shore standing and alongshore progressive linear shallow‐water modes. This inversion uses the Bayesian maximum a posteriori method (MAP). Infragravity edge waves on a moderately sloped ocean beach are characterized using 60 days of observations with low‐to‐moderate energy incident waves. Colocated P and UV sensors were deployed from the shoreline to 30 m depth, with an eight‐element PUV array in 7 m depth spanning 1.5 km alongshore. MAP estimates qualitatively agree with previous results obtained using the maximum likelihood estimator and smaller arrays. Edge waves average (28%, 14%, 43%) of the IG (P, U, V) variances in 7 m depth. Edge waves are most energetic at high tide when the shoreline slope is largest, possibly because the steep beach supports the multiple constructively interfering shoreline reflections required to form edge waves. The numerical wave model SWASH 1D assumes normally incident waves and neglects edge waves, but reproduces approximately the energy and cross‐shore structure of P and U.