This exploratory study was designed to identify factors implicating microbial influence on medicinal plant metabolomes. Utilizing a whole-microbiome approach, amplicon sequencing was used to identify the makeup of fungal and bacterial assemblages from endophytic (interior) and epiphytic (external) environments in two different sets of congeneric host-plant pairs, with collection of multiple samples of two medicinal plant species (Actaea racemosa, Rhodiola rosea) and two generic analogs (Actaea rubra, Rhodiola integrifolia). Diversity analysis of microbial assemblages revealed the influence of three primary factors driving variance in microbial community composition: host-plant taxonomy, the compartmentalization of microbial communities within discrete plant parts, and the scale of distance (microhabitat heterogeneity) between sampling locations. These three factors accounted for ~ 60% of variance within and between investigated microbiomes. Across all our collections, bacterial populations were more diverse than fungi (per compartment), and microbial density in epiphytic compartments (aerial parts, rhizosphere) were higher than those of endophytes (leaf and root). These comparative data point to key loci associated with variation between congeneric pairs and plant genera, providing insight into the complex and contrasting relationships found within this multi-kingdom coevolutionary relationship. Although reflective of only a limited set of botanical source materials, these data document the richness of a relatively unexplored component of the plant world and highlight the relevance of a whole-microbiome ecology-driven approach to botanical research and directed natural product investigations.