A high-performance local search engine that brings advanced RAG capabilities to your markdown knowledge bases and notes.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install qmd-2
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install qmd-2 using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is QMD - Quick Markdown Search?
QMD is a powerful local search engine designed specifically for markdown content, bridging the gap between traditional keyword-based search and modern semantic understanding. By integrating into the ecosystem of Openclaw Skills, it allows AI agents to instantly locate information across thousands of files using sophisticated retrieval methods like BM25 and vector embeddings. It serves as a privacy-focused, high-speed solution for developers and knowledge workers who need to query their personal documentation, journals, or project notes without relying on cloud-based search indexes.
QMD - Quick Markdown Search Use Cases
- Searching through a massive personal knowledge base for specific technical concepts like CAP theorem consistency.
- Finding documentation snippets when you only remember the context or intent but not the exact keywords.
- Integrating local documentation retrieval into AI agent workflows for better technical context and reduced hallucinations.
- Batch retrieving project notes using glob patterns for automated summary generation or code refactoring tasks.
How QMD - Quick Markdown Search Works
- Users add local directories containing markdown files to QMD collections to define the search scope.
- The system indexes content using both lexical (BM25) and vector embedding methods to enable hybrid search functionality.
- Search queries can be submitted via the CLI, MCP server, or HTTP API as simple text or complex structured objects.
- For complex queries, the engine uses HyDE (Hypothetical Document Embeddings) to generate an answer-like document to improve semantic matching.
- The engine performs reciprocal rank fusion (RRF) and reranking to deliver the most relevant snippets back to the Openclaw Skills interface.
QMD - Quick Markdown Search Setup
To get started with this skill, ensure you have Node.js installed and run the following commands:
# Install the QMD CLI globally
npm install -g @tobilu/qmd
# Add your markdown directory as a named collection
qmd collection add ~/my-markdown-docs --name docs
# Generate the search index and embeddings
qmd embed
QMD - Quick Markdown Search Data Schema & Taxonomy
QMD organizes information into collections and utilizes various search types to categorize and retrieve metadata. The data schema for queries is as follows:
| Feature | Type | Description |
|---|---|---|
| lex | Keyword | Uses BM25 for exact matches, code identifiers, and specific terms. |
| vec | Semantic | Uses vector embeddings for natural language questions and concepts. |
| hyde | Hypothetical | Generates a 50-100 word hypothetical answer to refine vector search. |
| intent | Signal | A steering signal used to disambiguate terms like performance or state. |
| collections | Filter | An array of strings used to limit search results to specific directories. |
QMD - Quick Markdown Search Advanced Features
- Hybrid retrieval combining BM25 and Vector scores with weighted fusion for maximum recall.
- HyDE support to improve retrieval quality on abstract or complex topics where keywords are unknown.
- Intent-driven disambiguation to steer AI agent results toward specific domains like web performance or team health.
- Multi-get functionality allowing for batch retrieval of documents using glob patterns or comma-separated lists.
- Full MCP server support, enabling seamless integration with any AI agent using Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Bitpanda
A comprehensive read-only bash CLI skill for querying Bitpanda account data including portfolio balances, trade history, and market prices.
Music Generator TopMediai
An automated AI music production tool for generating songs, background tracks, and lyrics via the TopMediai API.
Clawriosity
A personal curiosity engine that delivers daily science and culture insights through interactive quizzes, articles, and podcast scripts.
ZJTJ+SAR Four-Dimensional Strategy V2
An advanced A-share trading strategy that synthesizes main force capital tracking, trend analysis, and volume validation into a 100-point scoring system.
FTShare-cb-data
A comprehensive data skill set for querying Chinese convertible bond lists, fundamental information, and historical market trends.
Traffic Monitor Skill
A specialized network monitoring skill designed to track monthly bandwidth consumption and provide automated traffic alerts.
Featured*
