qmd

You can run local hybrid search across markdown notes and meeting transcripts by indexing your files on macOS or Linux, combining BM25 keyword search, vector search, and LLM reranking for fast on-device retrieval.

Yes, you can search personal knowledge bases entirely offline without cloud dependencies. The on-device indexing pipeline processes notes and docs locally, ensuring data privacy and fast retrieval.

The best way to combine keyword and semantic search locally is using a hybrid retrieval pipeline that applies BM25 for keyword matching, vector search for semantic similarity, and LLM reranking for high-quality results.

Yes, local vector search works with Hermes Agent workflows through MCP integration. You can expose search capabilities directly within your agent pipelines using the provided CLI and MCP interfaces.

You index documentation for on-device retrieval by installing the tool and pointing it at your collection of notes and docs. It builds a local index using BM25 and vector embeddings to enable fast searching.

macOS and Linux support offline knowledge base search with LLM reranking. These platforms can run the on-device indexing and hybrid retrieval pipeline to locate information across local documents.