Glycine receptors (GlyRs) mediate inhibitory neurotransmission in spinal cord motor and pain sensory neurons. Recent studies demonstrated apparently contradictory (potentiating versus inhibitory) effects of the endocannabinoid anandamide on these receptors. The present study characterised the effects of cannabinoid agonists on alpha1, alpha1beta, alpha2 and alpha3 GlyRs recombinantly expressed in HEK293 cells with the aims of reconciling effects of cannabinoids on these receptor subtypes and to establish the potential of different GlyR isoforms as novel physiological or analgesic targets for cannabinoids. The compounds investigated were anandamide, HU-210, HU-308, WIN55,212-2 and the endogenous non-cannabinoid, N-arachidonyl-glycine. The latter compound was chosen due to the structural similarity with anandamide and known analgesic actions in the spinal cord. Recombinant alpha1 and alpha1beta GlyRs were potentiated by anandamide and HU-210 at submicromolar concentrations, whereas WIN55,212-2 had no effect and HU-308 produced only weak inhibition. By contrast, N-arachidonyl-glycine exerted complex effects including both potentiation and inhibition. Anandamide had no effect at alpha2 or alpha3 GlyRs although the other cannabinoids produced potent inhibition. On alpha2 GlyRs, the inhibitory potency sequence was HU-210=WIN55,212-2>HU-308>N-arachidonyl-glycine but on alpha3 GlyRs it was HU-210=WIN55212=HU-308>N-arachidonyl-glycine. These results suggest that alpha1, alpha2 and alpha3 containing GlyRs exhibit distinct pharmacological profiles for cannabinoids. We conclude that cannabinoid agonists may be useful as pharmacological tools for selectively inhibiting alpha2 and alpha3 GlyRs. Our results also establish GlyRs as potential novel targets for endogenous and exogenous cannabinoids.