📚 Heimdall: Lightweight CPU-Only Memory Retrieval

While much of the ecosystem focuses on GPU acceleration and complex vector databases, this project takes a minimalist approach. It is a lightweight, CPU-only memory and ranked retrieval tool written primarily in JavaScript. The core philosophy here is simplicity and effectiveness. By dropping the requirement for expensive hardware, it provides a RAG (Retrieval-Augmented Generation) solution that can run on standard machines. The inclusion of MCP and semantic search topics indicates it is well-suited as a lightweight knowledge base plugin for local AI agents. If you just need to add basic memory to a local CLI assistant or a small agent without the overhead of deploying a heavy database, this compact, CPU-only approach is highly practical.