Continuous surveillance of composting processes would enable a feedback loop to be obtained for both analysis and process control. For this purpose, we designed e-noses based on a six-electrode quartz-crystal microbalance (QCM) array coated with affinity materials and molecularly imprinted polymers (MIP). They enable quantitative monitoring of volatile organic compounds (VOCs) emitted directly in a compost bin and are highly suitable tools for achieving on-line characterization of the degradation processes occurring. During grass and pine composting (duration 14 days and 40 days, respectively), we observed concentrations of up to 250 ppm of esters, 700 ppm of alcohols, 250 ppm of terpenes, and 90% relative humidity directly on-line with such a system and could validate the data off-line by GC-MS. The sensor also gave direct insight into the differences between the two composting batch types. Besides duration, during grass composting larger amounts of alcohols are emitted whereas relative amount of terpenes is twice as high for pine composting. Detailed correlation of the sensor and the GC-MS data allows approximate estimation of the sensitivity of the sensor materials towards analyte classes such as, e.g., aliphatic alcohols or terpenes.