Purpose: The rapidly expanding corpus of surgical literature presents an increasing challenge to researchers attempting to distil meaningful insights from masses of text. Text mining describes a technique for the automated extraction of information from textual data, analogous to the way numerical data is analysed. We developed MedVectors, a text mining platform for identifying associations, patterns, themes and specific data points across large volumes of unstructured surgical literature. Methodology: All 10,654 article abstracts from ANZ J Surg were pre-processed using Natural Language Processing (NLP) to produce machine-readable structured text. The transformer-based model BERTopic performed vectorisation and clustering, embedding in each term a representation of its meaning and associations. Results: MedVectors identified 300 unique topics that comprise the literary content of ANZ J Surg. We present these findings on interactive 2D topic maps and hierarchical tree diagrams. 10-year temporal analysis revealed the evolving foci of surgical literature, identifying trending and dissipating topics in the Australian surgical landscape. The MedVectors search algorithm offers a semantic-based article retrieval for improved accuracy and efficiency of surgical literature searching. Conclusion: Our analysis of ANZ J Surg demonstrates the computational advantages of text mining and AI for surgical literature exploration and synthesis. MedVectors provided unique insight into the literary content and the evolving thematic trends of the Australian surgical landscape.