khuedoan/qmd

QMD is a search engine that runs entirely on your own machine and helps you find things across all your scattered notes, meeting transcripts, documentation, and knowledge bases. Instead of digging through folders trying to remember where you wrote something down, you point it at your files and then search using everyday language. What makes it interesting is that it combines three different search approaches. First, it does traditional keyword matching, finding exact words like a standard search bar. Second, it does semantic search, understanding the meaning of your question to find related concepts even if the exact words don't appear. Third, it uses a local AI model to re-rank the best candidates, sorting the final results by true relevance. Everything runs locally, meaning your files never leave your computer and you don't need to pay for external AI API calls. The tool is built for people who work with AI assistants and agents. If you use something like Claude Desktop or Claude Code, you can connect this search engine directly to it so your AI assistant can search your personal notes and documents. A product manager could use it to instantly pull up old meeting notes about a specific project timeline, or a developer could use it to let their AI agent find relevant documentation before writing code. You can also use it as a standalone command-line tool just to find your own files. You can use it as a simple command-line tool, plug it into an AI assistant like Claude, or embed it directly into your own applications using its software library. The only notable requirement is that you need a relatively recent version of Node.js or Bun installed on your computer, and on macOS you need a specific version of the database software SQLite for certain features to work properly.