Water and air are generally treated as separate reservoirs of environmental DNA (eDNA) derived from the species resident in those respective environmental compartments. However, it is likely that eDNA routinely crosses the air–water boundary in both directions as a result of deposition, evaporation, or other processes. Here, we systematically tested methods of sampling eDNA at the air–water interface, showing for the first time that aquatic life can be consistently detected under standardized field conditions from passive air samples. We deployed four simple air samplers — three different kinds of filters and one open tray of deionized water — alongside paired water samples and visual counts over a six-week peak run of Coho salmon (Oncorhynchus kisutch) at a local spawning stream. We then quantified eDNA concentrations in both air and water (air: copies/cm2/day; water: copies/L) using quantitative PCR, to estimate (1) the concentration of target eDNA in air vs. water, and (2) the capture performance of each filter type. Passive air collectors captured quantitative airborne eDNA signals that covaried with salmon counts, despite air eDNA concentrations being approximately 25,000 times more dilute than water, although eDNA recovery varied with sampler design and orientation. We show the air–water interface can be a quantifiable source of aquatic genetic information in this system using simple, passive samplers that do not require electricity, making them appealing for biomonitoring in remote or resource-limited settings. This work points the way to using airborne eDNA as a promising pathway for biological information critical to conservation, resource management, and public-health protection.