Recent advances in food-web ecology highlight that most real food webs (1) represent an interplay between producer- and detritus-based webs and (2) are governed by consumers which are rampant omnivores; feeding on varied prey across trophic levels and resource channels. A possible avenue to unify these advances comes from models demonstrating that predators feeding on distinctly different channels may stabilize food webs. Empirical studies suggest many consumers engage in such behavior by feeding on prey items from both living-autotroph (green) and detritus-based (brown) webs, what we term “multichannel feeding,” yet we know little about how common such feeding is across systems and trophic levels, or its effect on system stability. Considering 23 empirical webs, we find that multichannel feeding is equally common across terrestrial, freshwater, and marine systems, with >50% of consumers classified as multichannel consumers. Multichannel feeding occurred most often at the first consumer level, indicating that most taxa at the herbivore/detritivore level are more aptly described as multichannel consumers, and that such feeding is not restricted to predators. We next developed a simple four-compartment nutrient cycling model for consumers eating both autotrophs and detritus with separate parameter sets to represent aquatic vs. terrestrial ecosystems. Modeling results showed that, across terrestrial and aquatic ecosystems, multichannel feeding is stabilizing at low attack rates on autotrophs or when attack rates are asymmetric (moderate on autotrophs while low on detritus), but destabilizing at high attack rates on autotrophs, compared to herbivory- or detritivory-only models. The set of conditions with stable webs with multichannel consumers is narrower, however, for aquatic systems, suggesting that multichannel feeding may generally be more stabilizing in terrestrial systems. Together, our results demonstrate that multichannel feeding is common across ecosystems and may be a stabilizing force in real webs that have consumers with low or asymmetric attack rates.