In this (very preliminary) project, I aim to develop a framework for studying narratives. I propose a novel methodology which leverages LLM capabilities to extract individual stories from a corpus of texts, creating a database of stories. Stories are then compared, via a notion of semantic similarity which respects story-structure, and agglomerated into narratives. Preliminary results attesting to the capabilities of the methodology in the context of inflation-related narratives will be furnished, along with proposed directions for future research. Accompanying preliminary theoretical conjectures, and the empirical hypotheses they generate, will also be discussed.