Bacterial vaginosis (BV) affects 20–30% of females and is associated with significant adverse outcomes, including increased susceptibility to sexually transmitted infections and a higher risk of preterm birth. It is characterised by a reduction in beneficial Lactobacillus species and a concurrent overgrowth of anaerobic bacteria such as Gardnerella vaginalis. This dysbiosis promotes the formation of polymicrobial biofilms on the vaginal epithelium, contributing to recurrent infections. Recurrence is further complicated by the broad-spectrum nature of current antimicrobials, such as clindamycin. We therefore aimed to employ a drug repurposing screen to identify novel pathogen-selective and antibiofilm molecules. Screening over 2,100 compounds identified several molecules with both inhibitory and antibiofilm activity against G. vaginalis but not against beneficial lactobacilli. Molecules of interest included an aminopeptidase N inhibitor, which inhibited biofilm metabolism at concentrations as low as 1.25 µM and demonstrated synergistic activity with the existing antimicrobial metronidazole (synergy score: 12.25). Furthermore, ebselen, an organoselenium compound with known anti-inflammatory activity, showed significant activity against metronidazole-resistant strains and exhibited antibiofilm activity against a polymicrobial, four-species BV biofilm model. Further analysis of structural ebselen analogues improved the therapeutic selectivity window against lactobacilli and demonstrated no cytotoxicity towards VK2 epithelial cells at concentrations up to 80 µM. To elucidate the cellular target of the compound, metatranscriptomic analysis was performed on the polymicrobial BV biofilm model at sub-inhibitory concentrations. These findings refine the list of lead compounds and reinforce drug repurposing as a viable strategy to address recurrent BV.