A local search engine designed for indexing and retrieving Markdown notes and documentation using keyword and semantic search.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install qmd-skill-3
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-skill-3 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 tool for developers and researchers who maintain extensive collections of Markdown files. It functions as a local search engine that allows for high-speed indexing and retrieval of documentation, notes, and personal knowledge bases. By integrating this capability into your workflow via Openclaw Skills, you can bridge the gap between static local files and your AI coding agent, enabling the agent to query your disk-bound knowledge with precision.
The tool is built on Bun and SQLite, utilizing BM25 for near-instant keyword matching and local GGUF models for advanced semantic vector searches. It is specifically optimized for Markdown, employing content-based chunking that handles even non-structured or messy notes effectively. This makes it an essential utility for anyone looking to build a retrieval-augmented generation (RAG) pipeline for their private local documents.
qmd (Quick Markdown Search) Use Cases
- Searching through extensive personal knowledge bases or developer logs stored in Markdown.
- Finding semantically related notes when keyword matches are insufficient.
- Retrieving specific documentation sections to provide context for AI agent tasks.
- Indexing local research repositories for fast, offline information retrieval.
- Automating the synchronization of search indexes for constantly evolving project notes.
How qmd (Quick Markdown Search) Works
- Add a local directory to the qmd system as a named collection using a file mask.
- Optional: Define a context for the collection to help the AI agent understand the nature of the indexed content.
- Execute the embedding process to generate vector representations of your files for semantic search.
- Perform searches using keyword matching (BM25) for speed or vector search for similarity.
- Retrieve full file contents or specific document chunks using unique IDs or file paths.
qmd (Quick Markdown Search) Setup
To get started with qmd as part of your Openclaw Skills toolkit, ensure you have Bun and SQLite installed, then run the following commands:
bun install -g https://github.com/tobi/qmd
qmd collection add /path/to/notes --name notes --mask "**/*.md"
qmd context add qmd://notes "My technical documentation"
qmd embed
qmd (Quick Markdown Search) Data Schema & Taxonomy
qmd manages data through structured collections and local vector stores.
| Component | Details |
|---|---|
| Collections | Named groups of files defined by a root path and a glob pattern (e.g., **/*.md). |
| Indexing | Uses SQLite with extensions for BM25 keyword search and vector storage. |
| Chunking | Content-based chunking (approx. 200-300 tokens) to ensure relevance during retrieval. |
| Models | Uses local GGUF models stored in ~/.cache/qmd/models/ for embeddings and reranking. |
qmd (Quick Markdown Search) Advanced Features
- Hybrid search mode that combines traditional keyword matching with LLM-based reranking for maximum accuracy.
- Support for cron-based automation to keep keyword and vector indexes fresh as files change.
- Agent-friendly output options including --json and --files flags for programmatic consumption.
- Local-first architecture ensuring all notes and embeddings stay on your machine without external API calls.
- Efficient multi-get commands for retrieving multiple documents simultaneously by path or ID.
SKILL.md
Loading
Related Openclaw Skills
Foundry
A self-evolving meta-extension that researches documentation and writes new tools, skills, and hooks to expand its own capabilities.
Feishu File Sender
An essential utility for AI agents to upload locally generated files and send them as messages via the Feishu OpenAPI.
Baidu Search
An AI-powered search tool that enables agents to query the Baidu Search Engine for live web data, documentation, and research topics.
Railway
A comprehensive tool for deploying and managing applications on the Railway cloud platform with zero configuration.
Claude Usage Calculator
A specialized tool to track and visualize Claude Max subscription consumption and rate limits from session data.
Discord Project Manager
An automated infrastructure for managing Discord-based project collaboration between multiple AI agents and human teams.
Featured*
