SmartCodeHub: LLM-Based Framework for Semantic Code Reuse in Reactive Programming

Abstract

Code reuse is essential for improving software productivity, yet developers still spend significant effort searching for and re-implementing similar code fragments. Existing snippet management tools rely primarily on keyword-based search, which fails to capture semantic relationships, particularly in reactive and asynchronous programming contexts. This paper presents SmartCodeHub, an AI-assisted snippet management framework that combines semantic code embedding, automated tag generation, and large language model (LLM) reasoning to support contextual code discovery and reuse. SmartCodeHub integrates a searchable snippet library with an interactive retrieval interface and cross-language support. Preliminary evaluation on JavaScript and Python projects indicates improvements in retrieval accuracy and reuse efficiency compared to conventional snippet tools. These early results suggest the feasibility of LLM-enhanced snippet ecosystems and highlight directions for a more broader and reproducible evaluation.