This thesis presents the design and implementation of a retrieval-augmented generation (RAG) system to organize and access scattered information about the city of Kiruna in northern Sweden. Because of long-term mining activity, Kiruna is being gradually relocated. Buildings, roads, and entire districts are being moved or rebuilt. As a result, a large amount of documentation has been produced over time. This includes planning reports, maps, images, municipal records, and research papers. These materials exist in many different places and formats, which makes them difficult to search and connect. General large language models are not trained on this local data, so they cannot reliably answer detailed questions about Kiruna’s transformation.