Objectives To develop a high throughput screening-compatible assay for the selection of species-specific antibiotics that do not harm human cells. Methods Staphylococcus aureus and human reporter cells continuously generating a fluorescence signal were competitively co-cultivated. The fluorescence signals were determined in the presence and absence of the specific antibiotic streptomycin and the toxic compound sodium azide. The results were compared with a standard cfu assay. Results In the absence of an effective antibiotic, S. aureus outgrew the human reporter cells and thus abolished the fluorescence signal. Conversely, the addition of streptomycin resulted in the growth of the reporter cells and a strong fluorescence signal. When sodium azide was added instead of streptomycin, only a very low background signal was obtained indicating toxicity and damage to the human reporter cells. The assay proved to be highly reliable (Z-factor >0.9) and high fluorescence signals correctly correlated with the efficient inhibition of S. aureus, as determined in comparative cfu assays. Conclusions In contrast to conventional cfu assays, the co-cultivation system allows the effects of a drug candidate on pathogens and human cells to be monitored simultaneously. Cytotoxic compounds can, therefore, be quickly ruled out during a primary screen. The nature of the screen also enables effective antibiotics to be identified without engineering the target pathogen to yield a fluorescence signal.