Balanced vegetarian diets are popular, although they are nearly absent in creatine and carnosine and contain considerably less carnitine than non-vegetarian diets. Few longitudinal intervention studies investigating the effect of a vegetarian diet on the availability of these compounds currently exist. We aimed to investigate the effect of transiently switching omnivores onto a vegetarian diet for 6 months on muscle and plasma creatine, carnitine and carnosine homeostasis. In a 6-month intervention, forty omnivorous women were ascribed to three groups: continued omnivorous diet (control, n 10), vegetarian diet without supplementation (Veg+Pla, n 15) and vegetarian diet combined with daily β-alanine (0·8–0·4 g/d) and creatine supplementation (1 g creatine monohydrate/d) (Veg+Suppl, n 15). Before (0 months; 0M), after 3 months (3M) and 6 months (6M), a fasted venous blood sample and 24-h urine was collected, and muscle carnosine content was determined by proton magnetic resonance spectroscopy (1H-MRS). Muscle biopsies were obtained at 0M and 3M. Plasma creatine and muscle total creatine content declined from 0M to 3M in Veg+Pla (P=0·013 and P=0·009, respectively), whereas plasma creatine increased from 0M in Veg+Suppl (P=0·004). None of the carnitine-related compounds in plasma or muscle showed a significant time×group interaction effect. 1H-MRS-determined muscle carnosine content was unchanged over 6M in control and Veg+Pla, but increased in Veg+Suppl in soleus (P<0·001) and gastrocnemius (P=0·001) muscle. To conclude, the body creatine pool declined over a 3-month vegetarian diet in omnivorous women, which was ameliorated when accompanied by low-dose dietary creatine supplementation. Carnitine and carnosine homeostasis was unaffected by a 3- or 6-month vegetarian diet, respectively.