Search local knowledge bases with hybrid BM25, vector, and reranking retrieval.
npx skills add https://github.com/ar0cket1/Hermes-Agent-Online-RL --skill qmd-ar0cket1
Please help me install this Agent Skill. Skill: qmd Source: https://github.com/ar0cket1/Hermes-Agent-Online-RL/tree/main/optional-skills/research/qmd Command: npx skills add https://github.com/ar0cket1/Hermes-Agent-Online-RL --skill qmd-ar0cket1
Local knowledge bases and notes are often scattered across folders and apps; qmd provides fast, private local search across markdown/text files using a hybrid retrieval engine.
Initialize your knowledge base by adding a collection with qmd collection add <path> --name <name> and index it with qmd embed.
High-intent search queries and answers about installing and using this skill.
You can search local notes and transcripts without cloud dependencies by using qmd to index markdown and text files locally, performing on-device hybrid retrieval via BM25, vector search, and LLM reranking.
Hybrid retrieval for local knowledge bases works by combining BM25 keyword search, vector search, and LLM-based reranking to deliver highly relevant, ranked results across multiple indexed collections.
Yes, you can search across multiple collections of meeting notes locally. qmd supports cross-collection search, allowing you to find every note mentioning a specific topic with unified ranked results.
Yes, you need a local Node.js environment to run qmd for vector search. After installing Node.js, you must index your documents and generate embeddings locally before executing searches.
The best way to find scattered project notes is using a local hybrid search engine like qmd, which indexes markdown and text files from various folders to provide fast, private search without cloud services.
winsorllcIndex documents with embeddings and search them by semantic similarity.
v36372Search the alpha-research vault using QMD commands.
jleechanorgIndex and search local knowledge bases with hybrid retrieval.
JuanMS20Index Markdown and text files for local hybrid search with BM25, vectors, and reranking.
