Deep Search Semantic Memory
Enhances Claude with local semantic search capabilities for oversized codebases and documentation using the qmd engine.
Deep Search (cm-deep-search) is an intelligent power-up for the CodyMaster framework that solves context window limitations by integrating local vector search via tobi/qmd. It automatically detects when a project has outgrown standard AI memory—typically exceeding 50 documentation files or 200 source files—and offers a zero-config bridge between active working memory and long-term semantic retrieval. By indexing your entire project locally, it allows Claude to find relevant historical specs, PRDs, and code patterns through semantic queries, ensuring the AI maintains high accuracy even in massive enterprise-scale repositories.