In vitro biofilm models are the basis for most studies of biofilm biology because they enable high-throughput analyses without the expenditure of animals. But how do we ensure that what we learn from in vitro studies is relevant in vivo? Biofilms grown in standard laboratory media do not interact with host factors and are thus profoundly different from in vivo biofilms. We therefore need in vitro models that are as in vivo-like as possible. We investigated how the addition of divalent cations and human plasma to brain heart infusion broth affected biofilm formation by Staphylococcus aureus and Staphylococcus epidermidis. We quantified the biofilm biomass and mechanical properties, imaged the 3D structure, and quantified antibiotic penetration and susceptibility. We also determined how the choice of plasma stabilizer affected biofilm growth and viability. Human plasma stimulated biofilm formation > 100 fold in different S. aureus strains, and increasing the calcium and magnesium concentration to physiological levels further doubled biofilm formation for MRSA. When grown in plasma, S. aureus biofilms consisted of fibrinogen-coated cell clusters interspersed in a matrix of fibrin fibers. This was in stark contrast to the homogenous biofilms formed in the laboratory media. While vancomycin susceptibility was similar for all biofilms, daptomycin more effectively eradicated biofilms grown in plasma. S. epidermidis biofilms were also strongly affected by human plasma. Strains that did not produce polysaccharides could not form biofilms in the absence of plasma, while in the presence of plasma, the structure of polysaccharide-producers and non-producers were indistinguishable. Extracellular polysaccharides may thus be of little consequence to S. epidermidis biofilms in vivo. Access to host factors is critical for the biofilm phenotype of both coagulase-positive and -negative staphylococci, and we urge that biofilms are routinely studied under in vivo-like conditions.