A local hybrid search engine designed to index and retrieve Markdown notes and documents with lightning speed.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install qmd-external
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-external 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?
qmd is a high-performance local search engine specifically built for Markdown-based knowledge bases and documentation. It bridges the gap between simple grep searches and complex vector databases by offering a hybrid approach. This Openclaw Skills utility enables users and AI agents to instantly find content within thousands of local files using BM25 keyword matching or deep semantic similarity via local GGUF models.
By running entirely on your local machine, qmd ensures your private notes and documentation stay secure while providing the retrieval performance required for modern AI workflows. It handles messy Markdown with ease, utilizing content-based chunking rather than requiring strict structural formatting. Whether you are managing a digital garden, a technical wiki, or personal journals, this tool provides the essential search infrastructure for your local workspace.
qmd Use Cases
- Searching personal knowledge bases, digital gardens, or vaults for specific information.
- Finding semantically related content when exact keyword matches are unavailable.
- Retrieving specific Markdown documents to provide context for AI agent tasks.
- Indexing local technical documentation for rapid developer reference.
How qmd Works
- The user defines a collection by pointing qmd to a local directory containing Markdown files.
- The system indexes the content, creating a searchable database optimized for BM25 keyword matching.
- Running the embed command triggers a local LLM to generate vector embeddings for semantic search capabilities.
- When a search is initiated, qmd scans the index and returns the most relevant chunks or files with associated relevance scores.
- Documents can be retrieved in full or as specific fragments via document IDs or file paths.
qmd Setup
To get started with this skill, ensure you have Bun and SQLite installed on your system. You can then install the tool and initialize your first collection using the following commands:
bun install -g https://github.com/tobi/qmd
qmd collection add /path/to/notes --name notes --mask "**/*.md"
qmd embed
Ensure your shell PATH includes $HOME/.bun/bin to allow Openclaw Skills to execute the binary correctly.
qmd Data Schema & Taxonomy
qmd organizes its data using a collection-based system and stores metadata in a local SQLite database. The following table describes the primary data elements:
| Element | Description |
|---|---|
| Collection | A logical grouping of files defined by a name and a glob pattern (e.g., 'notes'). |
| Doc ID | A unique identifier (prefixed with #) assigned to every indexed chunk. |
| BM25 Score | A metric used for keyword search relevance. |
| Vector Embedding | A numerical representation of text used for semantic (vsearch) lookups. |
| Model Cache | Local storage for GGUF models located at ~/.cache/qmd/models/. |
qmd Advanced Features
- Hybrid Querying: Combine keyword search with LLM reranking for the highest possible precision in search results.
- Automated Maintenance: Schedule re-indexing tasks using cron jobs to keep the search index synchronized with your file changes.
- Agent-Friendly Output: Utilize the --json and --files flags to generate machine-readable data for seamless integration with other Openclaw Skills.
- Multi-Get Retrieval: Efficiently fetch multiple documents at once using glob patterns or a list of document IDs.
SKILL.md
Loading
Related Openclaw Skills
Context Engineering Skill
A comprehensive routing framework for optimizing AI agent context, memory management, and multi-agent system design.
MoltMail Skill
Provides unique email addresses and full messaging capabilities for AI agents to communicate and collaborate.
MoltCredit Skill
A trust-based credit system enabling AI agents to extend credit lines, track balances, and settle debts via the X402 protocol.
ForkZoo
A GitHub-native digital pet system that uses AI to evolve and mutate your companion daily through automated workflows.
Rule Creation
Automatically formalizes and enforces AI behavioral constraints and policies through proactive Context Engineering.
Feynman Coach
An AI-powered personal tutor that uses the Feynman Technique to help you achieve deep mastery of any subject through interactive teaching and gap analysis.
Featured*
