Abstract: The Beijing Swift (Apus apus pekinensis) is a typical cavity-nesting bird that often nests inside holes and crevices in old architectures. Direct observation of their breeding behaviour is challenging and their breeding ecology is thus poorly studied. In this study, we analysed light-level geolocation data collected from six Beijing Swifts for the first time. Our results showed that geolocators can make comprehensive inference of their incubation period and behaviour. As a cost-effective and non-invasive method, geolocators can not only facilitate discovering migration routes, but also can be widely applied in the study of avian reproductive behaviour, especially in cavity-nesting bird species. We further discussed the characteristics and merits of this method and compared with other conventional nest-monitoring methods in recording birds.