Can I search multiple folders at once?

Yes, QMD allows you to organize files into collections and perform searches across specific groups or all indexed markdown documents simultaneously.

What is QMD in the context of Claude Code?

QMD is a skill that allows Claude to search your local markdown files and notes using both keyword-based and AI-powered semantic search methods via an MCP server or CLI.

Does QMD require an internet connection for searching?

QMD is designed to work with local markdown content. While it can use external embedding services, its primary function is to index and search your local files privately.

What are the different query types supported by QMD?

It supports 'lex' for keyword matching, 'vec' for semantic natural language questions, and 'hyde' for finding answers by generating a hypothetical response first.

How do I set up the QMD skill for Claude?

First, install the CLI via 'npm install -g @tobilu/qmd', add your notes folder with 'qmd collection add', and run 'qmd embed' to index your content.

QMD Quick Markdown Search

Name: QMD Quick Markdown Search
Author: kindship-ai

bykindship-ai

0•

学习与文档

Searches and retrieves information from local markdown knowledge bases and documentation using advanced lexical and semantic query methods.

QMD is a specialized search engine skill that enables Claude to interact with local markdown files, notes, and technical documentation. By utilizing a hybrid approach—combining BM25 lexical search for exact terms, vector-based semantic search for natural language queries, and Hypothetical Document Embeddings (HyDE)—it provides highly accurate information retrieval. This skill is ideal for developers managing large personal wikis or project-specific documentation who need Claude to reference local context quickly and precisely without manual searching.

主要功能

010 GitHub stars

02Query expansion using local LLMs to improve recall and disambiguate search intent

03Flexible retrieval via MCP tools for documents, snippets, and glob patterns

04Seamless integration with local embedding services for privacy-focused indexing

05Hybrid search combining lexical keywords (BM25) and semantic vector queries

06HyDE (Hypothetical Document Embeddings) support for finding answers via predicted document structure

使用场景

01Extracting specific code snippets or historical context from local changelogs using natural language

02Retrieving relevant notes from a personal Obsidian or markdown-based 'second brain' knowledge base

03Searching through large project documentation for specific API usage or architectural patterns

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add kindship-ai/kindship-qmd qmd

For use in Claude.ai and ChatGPT