Urban environments are thought to improve food security, by offering enhanced access to markets and income opportunities. Yet this idea is hard to test empirically due to an abundance of confounding factors and selection issues. This study leverages a resettlement program in China to provide the first quasi-experimental estimate of city life on food consumption and nutrition among low-income households. Lottery-determined timing of resettlement enables causal inference. We base our empirics on a 3-year panel and a range of difference-in-differences and matching methodologies. We find that those who were resettled to towns significantly increased both food consumption and diet variety, with increased intake of several macro- and micro-nutrients. Diet quality mostly improved, but we also found signs of over-consumption, notably of carbohydrates. Our evidence further suggests that our impacts are primarily due to improved market access. This stands in contrast to recent literature that finds little or no effect of living environments on food consumption. Instead, we reveal a significant impact of urban environments in shaping diets, bolstering the notion that supply-side channels do matter in some contexts.