The risk of contamination by cervical microbiota during transcervical sampling represents a fundamental methodological challenge in endometrial microbiome research. This study aimed to experimentally evaluate the efficacy of Pipelle and Endobrush endometrial sampling catheters in preventing this contamination. An in vitro cervix model with two anatomically distinct canal types (cylindrical and slit-like) was developed and filled with a synthetic cervical mucus containing a defined quantity of bacterial DNA. After catheter passage through the model cervical canal, a simulated ‘endometrial’ sample (sterile air) was collected and subjected to quantitative PCR analysis. Both catheter types facilitated substantial transfer of bacterial DNA from the cervical mucus into the endometrial sample. The median transfer of total bacterial DNA was 81.6% [54.4–107] for the Pipelle catheter and 29.8% [14.8–56.3] for the Endobrush catheter (p = 0.009), indicating that neither device provided sufficient protection for reliable characterization of the endometrial microbiota. Catheter efficacy was further dependent on cervical canal morphology and the specific microbial group analyzed. These findings demonstrate that transcervical sampling with either catheter type introduces a significant and variable degree of cervical contamination, thereby confounding the interpretation of endometrial microbiota data and underscoring the need to conceptualize and study a combined cervico-endometrial microbiota.