Food production is a primary contributor to climate change with greenhouse gas (GHG) emissions varying widely across food groups. In a randomized experiment, we examine the impact of providing individualized information on the GHG emissions of grocery purchases via a smartphone app, compared to providing information on spending. Carbon footprint information decreases GHG emissions from groceries by an estimated 27% in the first month of treatment, with an estimated 45% reduction in emissions from beef, the highest emissions food group. Treatment effects fade in the longer-run along with app engagement. However, we find evidence of persistent effects among those who remain engaged with the app. Our results suggest that individualized carbon footprint information can reduce the climate impact of food consumption but requires sustained engagement.