Trophic interactions can be both ephemeral and difficult to document, rendering their sampling often incomplete and context-dependent, which makes construction, analysis, and comparison of food webs challenging. Biological traits are central in determining co-occurrence of species (through dispersal, environmental, and interaction filters), as well as the potential for species interactions (through trait matching). Thereby, supplementing empirical, taxonomy-based information on trophic links with trait-based inference may help us build more realistic and adaptable food webs. Here, we go beyond taxonomy to document (i) how traits (e.g., body size, metabolic category and feeding strategy) contribute to local food web structure, and (ii) how associations of consumer-resource traits are structured. We built a trophic-link based trait-interaction network—or trait web—by combining multivariate approaches and network analysis. We found that consumer-resource associations organize into trait profiles that reflect the general vertical structure of the food web, as well as identify groups of limited sets of highly interacting traits. Finally, we discuss the implications of the findings for generating comprehensive and adaptive food webs.