Does my documentation data leave my local machine?

No, QMD is designed to be a local search engine. Indexing and searching happen on your device, ensuring your private notes and internal documentation remain secure.

What is hybrid search in QMD?

Hybrid search combines traditional keyword matching (BM25) with vector-based semantic search. This means Claude can find documents even if you don't use the exact words contained in the file.

What is the QMD Claude Code Skill?

QMD is a local-first search engine skill that allows Claude to index and search through your markdown files, notes, and documentation using both keyword and semantic matching.

How do I install the necessary components for this skill?

The skill requires the qmd CLI or MCP server. You can install it via Bun using the command: bun install -g https://github.com/tobi/qmd.

When should I use 'search' vs 'vsearch' vs 'query'?

Use 'search' for fast keyword matches, 'vsearch' for conceptual/semantic matches, and 'query' for the highest quality hybrid results where speed is less critical.

QMD Markdown Knowledge Search

Name: QMD Markdown Knowledge Search
Author: tobi

bytobi

•

5,804

Content Management

Searches personal markdown knowledge bases and documentation using a local hybrid search engine combining BM25 and vector semantic search.

About

QMD is a specialized skill for Claude Code that enables lightning-fast, local-first search across your personal markdown notes, meeting transcripts, and project documentation. By leveraging a hybrid search approach—combining traditional keyword matching with semantic vector search and LLM-powered re-ranking—it allows users to find information using natural language queries even when exact keywords aren't present. It is an essential tool for developers and knowledge workers who need to instantly retrieve context from extensive local documentation or personal knowledge bases while maintaining data privacy.

Key Features

Multi-collection management to organize notes, transcripts, and docs separately
Local-first execution for data privacy and high-speed indexing
5,804 GitHub stars
Hybrid search combining BM25 keyword matching and semantic vector search
Support for document retrieval via doc IDs, file paths, or glob patterns
LLM-powered result re-ranking for maximum search accuracy

Use Cases

Searching personal knowledge bases to synthesize information for research tasks
Finding specific information across thousands of meeting notes or transcripts
Retrieving internal project documentation when exact keywords are forgotten

About

Key Features

Multi-collection management to organize notes, transcripts, and docs separately
Local-first execution for data privacy and high-speed indexing
5,804 GitHub stars
Hybrid search combining BM25 keyword matching and semantic vector search
Support for document retrieval via doc IDs, file paths, or glob patterns
LLM-powered result re-ranking for maximum search accuracy

Use Cases

Searching personal knowledge bases to synthesize information for research tasks
Finding specific information across thousands of meeting notes or transcripts
Retrieving internal project documentation when exact keywords are forgotten