Diet is a key determinant of human and planetary health, but accurately measuring dietary intake remains challenging. Traditional self-reporting tools are imprecise, compromising our ability to accurately link diets with health outcomes. Modern technologies, including smartphone apps, image-based methods and biomarkers of food intake (BFIs), offer promise but bring their own caveats. App- and image-based methods reduce bias and reporting burden, but remain partly self-reported, and are thus prone to errors similar to those of traditional methods. Omics-based BFIs (that is, metabolites, food-related DNA or food proteins) are objective measures derived from biological samples; however, they mostly reflect recent intake, and require careful sampling alignment to estimate habitual diets. Here we discuss the drawbacks and opportunities for all dietary tools and propose strategies to integrate technologies along with multisampling for longitudinal measurements, for a new era in dietary assessment that can clarify the impact of diets, dietary components and dietary behaviour on human and planetary health.