Simple File Vector Store

Summary

This MCP server, developed by lishenxydlgzs, provides a vector store for local files and directories, enabling semantic search across their contents. Built with TypeScript and leveraging libraries like LangChain and FAISS, it offers tools for initializing the store and performing searches. The implementation focuses on creating embeddings from supported file types (Markdown, HTML, JSON, plain text) using the all-MiniLM-L6-v2 model and storing them efficiently. It's particularly useful for teams wanting to add semantic search capabilities to their local document repositories, enabling use cases such as context-aware information retrieval, automated document linking, and AI-assisted content exploration without relying on external search services.