DSH-RAG

In our daily work, we often need to query or retrieve information from locally stored documents, but traditional chat tools usually can't access these files directly or rely on the model's decisions for retrieval, leading to unreliable results. DSH-RAG solves this by converting authorized local folders into searchable knowledge bases. It uses explicit /rag commands to complete retrieval before the model answers, providing verifiable evidence with source citations. Its core capabilities include managing named knowledge bases, session-scoped search, hybrid retrieval with offline fallback, and incremental parsing of local documents. This plugin is ideal for researchers, writers, IT professionals, and anyone who needs to efficiently manage and retrieve information from large collections of local documents, all while prioritizing data privacy with local processing.