qmd

To search local markdown notes and meeting transcripts offline, you can index your local collections and query them directly. This allows you to retrieve text from your personal knowledge base without requiring an internet connection.

Hybrid search for personal knowledge bases combines BM25 keyword matching, vector search, and LLM reranking. This mechanism blends exact term retrieval with semantic understanding to deliver highly accurate document and transcript results.

Yes, local search works with MCP integration to connect Hermes Agent workflows. You can configure your indexed collections to be queried directly through the Model Context Protocol, enabling agents to retrieve notes and documents seamlessly.

Yes, you can use CLI to query your project documentation locally. By configuring your collections and running commands, you can perform fast offline searches across your indexed markdown notes and project docs.

The best way to retrieve information from meeting transcripts is using a hybrid search pipeline that combines BM25, vector search, and LLM reranking. This approach ensures accurate local retrieval by blending keyword and semantic search.

You use LLM reranking for local document search to improve result accuracy after initial retrieval. By applying LLM reranking to vector and BM25 results, you get higher quality matches from your personal knowledge base.