Forest canopies are highly diverse ecosystems, but despite several decades of intense research, there remain substantial gaps in our knowledge of their biodiversity and ecological interactions. One fundamental challenge in canopy research is the limited accessibility of the ecosystem. Consequently, previous studies have relied on the application of either highly invasive methods such as chemical knockdown, or on time-consuming and expensive setups such as canopy walkways or cranes. Therefore, time- and cost-efficient, ideally minimally invasive yet comprehensive applications are required to help close this knowledge gap. High-throughput metabarcoding of environmental DNA (eDNA) collected from water, soil, or air provides a minimally invasive method for biodiversity assessment, yet its potential for canopy biodiversity monitoring has not been explored. Herein, we conducted metabarcoding of eDNA washed off the canopy via rainwater to explore its potential for biodiversity monitoring and ecological research. We placed four 1 m² rain samplers beneath the canopies of four different trees (beech, oak, larch, and pine) prior to a major rain event, filtered eDNA from the collected rainwater, and performed cytochrome c oxidase subunit I (COI) gene metabarcoding to profile the invertebrate community. Additionally, we collected and identified all specimens present in the rainwater to assess if eDNA only came from specimens physically present in the rainwater. We detected 50 invertebrate species by eDNA metabarcoding, of which 43 were not physically present in the water sample, thus likely representing true canopy biodiversity signals. Furthermore, we observed distinct species occurrence patterns corresponding to the four trees, suggesting that ecological patterns such as host specificity can potentially be assessed using the method. In conclusion, our study provides a proof of principle that rainwash eDNA metabarcoding offers a minimally invasive and comprehensive method for biodiversity monitoring in tree canopies.