RCualiText: An Open-Source R/ShinyWeb Application for Qualitative Analysis

Qualitative researchers often face a practical bottleneck: large volumes of unstructured text and few affordable tools to work through it. Commercial CAQDAS packages exist, but their licensing costsand steep learning curves shut out many scholars, particularly those at resource-constrained institutions. We developed RCualiText, a free, open-source R/Shiny application, to fill that gap. The toollets users import documents, assign codes to selected passages, build hierarchical code structures, retrieve coded segments, and produce frequency charts and co-occurrence networks. What sets itapart from existing open-source alternatives is a semantic analysis layer built on OpenAI’s GPT-4.1 and text-embedding-3-small models. Coded fragments can be clustered by meaning, cross-code similarities flagged, and relationships plotted in two dimensions through t-SNE, UMAP, or PCA. The application also provides LLM-based coding validation, coherence scoring, and AI-assisted reportdrafting. We tested it with simulated interview data. RCualiText handled multiple codes without difficulty, kept full audit trails, and produced outputs comparable to commercial tools, while venturing into semantic territory that proprietary packages have not yet reached.