wiki-index-qmd

You can index semantic chunks for local research by placing properly formatted JSON files into the data/chunks/ directory and using the QMD CLI to boot the indexing process. This builds a local collection for private retrieval.

Hybrid search for article retrieval combines BM25 keyword matching with vector similarity to find relevant semantic chunks. This approach improves topic exploration by balancing exact term matching with contextual meaning.

Yes, local RAG indexing works offline for private data handling by keeping all chunks and search processes on your local machine. It requires no external network calls, ensuring data privacy during retrieval and iteration.

Yes, you need the QMD CLI installed to boot the indexing and search processes. The workflow relies on the CLI to read chunks from data/chunks/ and build the searchable local collection.

To set up JSON chunks for local indexing, format your semantic chunks as proper JSON files and place them in the data/chunks/ directory. The QMD CLI will read this directory to index the chunks into the collection.

Yes, you can use local semantic search to prototype and refine agent prompts offline before production. The hybrid search retrieval allows rapid iteration on RAG queries without requiring cloud infrastructure.