The purpose of this study was to evaluate intrafraction prostate motion in patients treated with proton therapy and an endorectal balloon or a hydrogel spacer using orthogonal x-rays acquired before and after each treatment field. This study evaluated 10 patients (662 fields throughout treatment) treated daily with an endorectal balloon (ERB) and 16 patients (840 fields throughout treatment) treated with a hydrogel spacer (GEL) without an ERB. Patient shifts were recorded before and after each treatment field, correlated with a treatment time, using x-ray imaging and implanted fiducial alignment. For each shift, recorded in X, Y, and Z, a 3D vector was calculated to determine the positional change. There was a statistically significant difference in the mean vector shift between ERB (0.06 cm) and GEL (0.09 cm), (P < 0.001). The mean includes a large number of zero shifts, but the smallest non-zero shift recorded was 0.2 cm. The largest shifts were, on average, in the Z direction (anterior to posterior). The average Z shift was +0.02 cm for both ERB and GEL. There was no statistical difference between ERB and GEL for shifts greater than 0.3 cm (P = 0.13) or greater than 0.5 cm (P = 0.36). For treatment times between 5 and 9 min, a majority of shifts were less than 0.2 cm, 85.9% for ERB and 73.2% for GEL. There was a significant positive correlation between the vector shifts and field time for both ERB (r = 0.2, P < 0.001) and GEL (r = 0.07, P < 0.04). We have shown that prostate motion is clinically comparable between an ERB and a hydrogel spacer, and the time dependencies are similar. A large majority of shifts for both ERB and hydrogel are well within a typical robust planning margin. For GEL patients, we chose to maintain slightly larger planning margins than for ERB due to already improved rectal sparing with GEL.