Player not loading? Watch on YouTube
The tutorial explains the speaker's "load and layer" method for giving Claude context from a second brain. Their example uses Obsidian to view a folder of Markdown files on their computer. An AI agent named Chad searches those notes before helping with member onboarding and workshop preparation. The speaker says this reduces how often they need to explain their business.
QMD handles context loading. The speaker describes it as a free GitHub tool that searches Markdown notes by meaning rather than relying on filenames or exact wording. They recommend using Claude Cowork or Claude Code to help set it up, then adding instructions to a Claude file to query QMD before the first response and again when uncertainty arises. This is a workflow demonstration, not a measured test of retrieval accuracy.
Context layering adds the relevant business area, project, conversation thread and connected notes. It requires a map that explains how the knowledge base fits together. The speaker uses layering alongside QMD and suggests it for connected systems such as Notion or Tana that lack the semantic search described here.
The setup depends on recording useful business knowledge and making it accessible to Claude. Local note storage alone does not establish that the full workflow runs offline.
