Artificial night-time illumination of natural habitats has increased dramatically over the past few decades. Generally, studies that assess the impact of artificial light on various species in the wild make use of existing illumination and are therefore correlative. Moreover, studies mostly focus on short-term consequences at the individual level, rather than long-term consequences at the population and community level—thereby ignoring possible unknown cascading effects in ecosystems. The recent change to LED lighting has opened up the exciting possibility to use light with a custom spectral composition, thereby potentially reducing the negative impact of artificial light. We describe here a large-scale, ecosystem-wide study where we experimentally illuminate forest-edge habitat with different spectral composition, replicated eight times. Monitoring of species is being performed according to rigid protocols, in part using a citizen-science-based approach, and automated where possible. Simultaneously, we specifically look at alterations in behaviour, such as changes in activity, and daily and seasonal timing. In our set-up, we have so far observed that experimental lights facilitate foraging activity of pipistrelle bats, suppress activity of wood mice and have effects on birds at the community level, which vary with spectral composition. Thus far, we have not observed effects on moth populations, but these and many other effects may surface only after a longer period of time.