Diets can be consistent with planetary limits and health targets at the individual level.

Progress towards eliminating hunger and promoting sustainable diets is lagging, with food systems damaging ecosystems and over 700 million people undernourished. Here we develop a linear programming model that identifies food combinations that satisfy both environmental and nutritional constraints. Using US-specific data, the model considers the environmental and nutritional characteristics of more than 2,500 food items consumed in the USA, optimizing diets based on the healthy life gained from the Health Nutritional Index. Aligned with the Paris Agreement's 1.5 °C target, various diets are found to offer up to 700 min of healthy life gained per week, while reducing climate impacts by a factor of seven. Vegan, vegetarian and flexitarian diets that limit meat consumption to 255 g per week (pork and poultry) best met environmental and nutritional constraints. Grains, legumes and nuts were the primary protein sources. These diets provide a range of specific options for consumers and actionable targets for policy recommendations.