Background: Food webs form the basis of biological communities, though empirical research has been hindered by difficulties in quantifying interactions. Metabarcoding from predator gut content extractions with universal primers promises to provide simple and rapid insights into food web interactions. However, the highly overabundant predator DNA often completely outcompetes that of the digested prey DNA during PCR, impeding the ability to assess the abundance and diversity of prey items. Methods: Focusing on the issue of overabundance of predator DNA amplified by a commonly used COI primer pair, we use predator lineage-specific SNPs at the 3’-end of PCR primers to selectively block out predators from amplification. While this approach largely prevents predator amplification, it retains high taxonomic versatility for prey lineages. We introduce a novel multilocus assay, targeting four nuclear and mitochondrial rDNA markers and test our approach in a diverse set of spiders from 12 families. We estimate the recovered prey DNA proportions and compare the taxonomic composition of prey communities between markers. Using a feeding experiment, we also explore recovery of prey DNA over time. Results: While commonly used COI primers yield low and very unpredictable amounts of prey DNA, our assay allows for a considerable and consistent prey enrichment across all tested species. The recovered prey’s taxonomic composition is comparable between markers and supports results acquired by COI. The new marker set can be amplified in a simple multiplex PCR, considerably reducing the necessary workload. Conclusions: Our multi-locus approach allows the generation of an unprecedented amount of prey data at low cost and effort. Lineage specific PCR is taxonomically versatile and could readily be adapted to any prey-predator interaction, opening up the opportunity for community-wide studies on food web interactions.