Glucagon-like peptide-1 receptor (GLP-1R) agonists are effective treatments for type 2 diabetes and obesity. We recently described biased peptide GLP-1R agonists modelled on exendin-4 which uncouple the pronounced endocytosis that usually accompanies GLP-1R activation, leading to prolongation of intracellular signalling responses. Here, we show that the metabolic consequences of biased GLP-1R activation in vivo are dominated by improvements in blood glucose, without concomitant increases in their anorectic properties (40-fold relative preference for glycaemic vs. anorectic effects). To investigate this disparity, we have compared cell type-specific responses to biased GLP-1R agonists using in vitro beta cell and neuronal models, in case the downstream manifestations of bias differ according to the tissue in which they act. We also delivered these ligands directly into the CNS in mice to bypass the blood-brain-barrier, under which circumstances the anticipated efficacy increase of the biased agonist was restored. Biodistribution studies using the same compounds conjugated to near-infrared fluorophores, which allow deep tissue imaging of optically cleared intact brain and pancreas specimens, have also provided a means to assess whether biased ligands have equal or differential access to appetite regulatory centres in the CNS.