Background <p>The transoral endoscopic thyroidectomy vestibular approach (TOETVA) utilizes an intraoral incision, raising concerns about surgical-site infections (SSI) originating from the oral cavity. The types and sources of bacteria that may cause TOETVA infection are not clear, and the optimal antibiotic prevention plan is still unclear, lacking targeted perioperative management.</p> Materials and methods <p>In this prospective, non-randomized controlled experimental study, drainage-fluid from 52 TOETVA and 15 open surgery patients (March – December 2025) was analyzed using 16S rDNA sequencing and clinical review. TOETVA patients were allocated to three perioperative antibiotic regimens: preoperative cefuroxime (Group A, <i>n</i> = 18), postoperative cefuroxime (Group B, <i>n</i> = 16), or postoperative metronidazole (Group C, <i>n</i> = 18). Inclusion criteria were initial surgery for T1N0M0 papillary thyroid carcinoma, no prior neck radiation, and patient preference for a scar-free neck. The primary outcome was postoperative SSI rate. Secondary outcomes included changes in drainage-fluid bacterial abundance, pus culture results, effects of different oral cleaning protocols, and microbial correlation between drainage-fluid and disinfected neck skin or oral vestibule. Statistical analyses employed the Chi-Square test, Kruskal-Wallis test, and Spearman Correlation.</p> Results <p>No SSI occurred in cefuroxime group, while two infections were recorded in the metronidazole group. Analysis of the top 30 differentially abundant bacterial species showed that cefuroxime reduced the abundance of 15/30 (50%) within 48&#xa0;h, compared to only 4/30 (13.3%) with metronidazole (OR: 6.50, 95% CI: 1.90-22.26, <i>P</i> = 0.0023). Microbial composition in drainage-fluid strongly correlated with povidone-iodine-disinfected chin skin (ρ = 0.8195, 95% CI: 0.6865–0.8994, <i>P</i> &lt; 0.001). However, pus cultures from infected cases grew oral streptococci, not skin flora. Among four oral cleaning regimens, povidone-iodine most effectively reduced streptococcal load, whereas tinidazole better preserved the overall microbiota balance.</p> Conclusions <p>Although the drainage-fluid following TOETVA primarily reflects skin microbiota, clinical surgical-site infections are mainly attributed to oral streptococci. Cefuroxime provides superior perioperative antibacterial coverage compared to metronidazole. Based on clinical practice considerations, a comprehensive oral management process can be helpful for the treatment of TOETVA.</p> Trial registration <p>Retrospectively registered in Chinese Clinical Trial Registry on April 8, 2026 (Registration number: ChiCTR2600122005).</p>