To cause meningitis, bacteria move from the bloodstream to the brain, crossing the endothelial cells of the blood-brain barrier. Most studies on how bacteria cross the blood-brain barrier have been performed in vitro using cultured endothelial cells, due to a paucity of animal models. Group B Streptococcus (GBS) is the leading cause of bacterial meningitis in neonates and is primarily thought to cross the blood-brain barrier by transcytosis through endothelial cells. To test this hypothesis in vivo, we used optically transparent zebrafish larvae. Time-lapse confocal microscopy revealed that GBS forms extracellular microcolonies in brain blood vessels and causes perforation and lysis of blood-brain barrier endothelial cells, which promotes bacterial entry into the brain. Vessels infected with GBS microcolonies were distorted and contained thrombi. Inhibition of clotting worsened brain invasion, suggesting a host-protective role for thrombi. The GBS lysin cylE, implicated in brain invasion in vitro, was found dispensable in vivo. Instead, pro-inflammatory mediators associated with endothelial cell damage and blood-brain barrier breakdown were specifically upregulated in the zebrafish head upon GBS entry into the brain. Therefore, GBS crosses the blood-brain barrier in vivo not by transcytosis, but by endothelial cell lysis and death. Given that we observe the same invasion route for a meningitis-associated strain of Streptococcus pneumoniae, our findings suggest that streptococcal infection of brain blood vessels triggers endothelial cell inflammation and lysis, thereby facilitating brain invasion.