RGW is a lightweight R-language implementation of the affine-invariant Markov Chain Monte Carlo sampling method of Goodman & Weare (2010). The implementation is based on the description of the python package emcee (ascl:1303.002).