Sambalpuri Sarees in the Digital Loom: Metadata Schema and Natural Language Query Interface Design

Sambalpuri sarees, traditional handloom textiles from Odisha, India, represent significant cultural heritage through their distinctive bandha (ikat) weaving techniques and symbolic motifs. Despite their historical and artistic value, digital documentation of these textiles remains limited due to the absence of domain-specific metadata structures and accessible retrieval mechanisms. This study proposes a customized metadata schema that extends the Dublin Core standard with textile-focused elements designed for Sambalpuri saree description. The schema is implemented within an open-source DSpace repository and integrated with a natural language query interface using a Model Context Protocol (MCP) architecture, where a language model functions as an interpretive layer between user queries and structured metadata. The system supports multilingual access and facilitates semantically guided discovery while maintaining metadata-driven retrieval. Qualitative evaluation with a curated dataset demonstrates successful schema implementation, feasible natural language interaction, and improved accessibility for diverse user groups. The study presents a lightweight, repository-oriented approach that contributes to the sustainable documentation and dissemination of intangible textile heritage.