The complexity underlying protein-protein interaction (PPI) networks calls for the development of comprehensive knowledge bases organizing PPI-related data. The constant growth and high reliability of structural data make them a suitable source of evidence for the determination of PPI. We present LEVELNET, a fully-automated and scalable environment designed to integrate, explore, and infer protein interactions and non-interactions based on physical contacts and other PPI sources, including user-defined annotations. LEVELNET helps to break down the complexity of PPI networks by representing them as multi-layered graphs and allowing the selection of subnetworks and their direct comparison. LEVELNET proposes an interactive visualisation based on a user-friendly web interface. LEVELNET applications are multiple. It allows to explore PPIs of biological processes, identify co-localised partners, assess PPI predictions from computational or experimental sources, unravel cross-interactions, show and compare multiple PPI sources, and help creating PPI benchmarks with specific properties. Availability: LEVELNET is freely available to the community at http://www.lcqb.upmc.fr/levelnet/