Setting up QMD for my wiki search

Andrej Karpathy recently posted about pattern for building personal knowledge bases using LLMs.

The main focus point is something else but he mentioned about QMD as a search engine on top of the wikis, docs etc.

An on-device search engine for everything you need to remember. Index your markdown notes, meeting transcripts, documentation, and knowledge bases. Search with keywords or natural language. Ideal for your agentic flows.

QMD combines BM25 full-text search, vector semantic search, and LLM re-ranking—all running locally via node-llama-cpp with GGUF models.

I have been maintaining my wiki since 2019. It has a bunch of docs, wikis, notes, daily diaries etc.

Graph view of my wiki

The biggest cluster is because of my daily notes, one new markdown doc every day.

Daily diary notes

I have setup QMD for my wiki collection along with separate collection of LLM generated knowledge docs etc.

QMD index, embedding model and re-ranking model

I can now use QMD MCP to ask questions like below.

Using Claude Code to query my wiki using QMD MCP