Quick Summary: 📝
QMD is a local, on-device search engine designed to index and query personal documents, notes, and knowledge bases. It combines traditional keyword search, semantic vector search, and LLM re-ranking to provide high-quality, context-aware results, making it ideal for integration with AI agents.
Key Takeaways: 💡
✅ QMD is an on-device, private search engine for all your markdown notes, documents, and knowledge bases.
✅ It combines keyword (BM25), semantic (vector), and LLM re-ranking for highly accurate and relevant search results.
✅ All processing runs locally using GGUF models, ensuring data privacy, speed, and no cloud dependency.
✅ The unique 'context' feature provides LLMs with deeper understanding, significantly enhancing AI agent performance.
✅ QMD is ideal for empowering AI agents with rich, structured, and contextual access to your personal knowledge.
Project Statistics: 📊
- ⭐ Stars: 26086
- 🍴 Forks: 1638
- ❗ Open Issues: 47
Tech Stack: 💻
- ✅ TypeScript
Have you ever found yourself sifting through countless markdown notes, meeting transcripts, or documentation, desperately searching for that one crucial piece of information? It's a common developer dilemma, and it often feels like trying to find a needle in a digital haystack. This is exactly where QMD, or Query Markup Documents, steps in as a game-changer. Imagine having a personal, lightning-fast search engine right on your device that understands what you mean, not just the keywords you type. That's QMD in a nutshell – it indexes all your valuable text-based knowledge, making it instantly retrievable. It's designed to be your on-device brain for everything you need to remember.
What makes QMD truly stand out is its intelligent approach to search. It doesn't just rely on one method; it cleverly combines three powerful techniques to give you the best possible results. First, there's the traditional full-text keyword search (BM25), which is super fast for direct matches. Second, it employs vector semantic search, meaning it understands the meaning behind your query, even if you use different words. So, if you search for "deployment steps," it might find documents discussing "how to put code into production." Finally, it uses an advanced LLM re-ranking process, which is like having a smart assistant refine the search results to surface the most relevant information for your specific need. The truly exciting part? All of this sophisticated processing happens entirely on your local machine, powered by node-llama-cpp and efficient GGUF models. This means your data stays private, your searches are incredibly fast, and you don't need an internet connection or expensive cloud APIs.
Getting started with QMD is remarkably straightforward. You can easily create different 'collections' for your various types of documents, whether they're personal notes, work documentation, or meeting summaries. One of QMD's most innovative features is its 'context' system. You can add descriptive context to each collection, like "Personal notes and ideas" for your notes collection. When QMD retrieves information from these collections, it also provides this context. This is incredibly powerful, especially when integrating with AI agents. By giving the agent not just the raw document snippet but also the context of what that document represents, LLMs can make much more informed and accurate decisions, leading to significantly better agentic workflows.
Developers will find QMD invaluable for a multitude of reasons. Beyond just organizing personal knowledge, it's a powerful tool for building more intelligent AI agents. The ability to export search results in structured JSON or as lists of files, coupled with its native support for agentic outputs, makes it a perfect fit for giving your AI agents access to a rich, private knowledge base. Whether you're building a chatbot that needs to pull specific details from your internal documentation or an automation script that requires contextual information from past meetings, QMD provides the backbone. Furthermore, for those who need even tighter integration, QMD offers an MCP (Model Context Protocol) server, allowing for persistent connections and even simpler setup with tools like Claude Desktop, or directly via command line.
The shift towards local-first AI tools is gaining momentum, and QMD is at the forefront of this movement. By keeping your search and reasoning capabilities entirely on your device, it offers unparalleled privacy, speed, and cost-effectiveness. No more worrying about data leaving your machine or racking up API bills. It empowers you to truly own and leverage your personal and professional knowledge in a way that was previously complex or expensive. If you're looking to supercharge your personal knowledge management or build more robust, private AI agents, QMD is definitely a project you'll want to explore.
Learn More: 🔗
🌟 Stay Connected with GitHub Open Source!
📱 Join us on Telegram
Get daily updates on the best open-source projects
GitHub Open Source👥 Follow us on Facebook
Connect with our community and never miss a discovery
GitHub Open Source
For further actions, you may consider blocking this person and/or reporting abuse
