We measured stacked weak lensing cluster masses for a sample of 1323 galaxy clusters detected by the RedGOLD algorithm in the Canada-France-Hawaii Telescope Legacy Survey W1 and the Next Generation Virgo Cluster Survey at 0.2 < z < 0.5, in the optical richness range 10 < lambda< 70. This is the most comprehensive lensing study of a similar to 100% complete and similar to 80% pure optical cluster catalog in this redshift range. We test different mass models, and our final model includes a basic halo model with a Navarro Frenk and White profile, as well as correction terms that take into account cluster miscentering, non-weak shear, the two-halo term, the contribution of the Brightest Cluster Galaxy, and an a posteriori correction for the intrinsic scatter in the mass-richness relation. With this model, we obtain a mass-richness relation of log M-200/M-circle dot = (14.46 +/- 0.02)+(1.04 +/- 0.09) log(lambda/40) (statistical uncertainties). This result is consistent with other published lensing mass-richness relations. We give the coefficients of the scaling relations between the lensing mass and X-ray mass proxies, L-X and T-X, and compare them with previous results. When compared to X-ray masses and mass proxies, our results are in agreement with most previous results and simulations, and consistent with the expected deviations from self-similarity.