RAG Context

This MCP server provides persistent memory and context management for AI assistants using local vector storage and SQLite database, enabling semantic search and indexed retrieval of stored information. Built with TypeScript using Vectra for vector similarity search and Xenova/all-MiniLM-L6-v2 for local text embeddings, it offers two core tools: setContext for storing information with automatic vectorization and metadata support, and getContext for retrieving relevant context through hybrid semantic search with configurable similarity thresholds. The implementation runs entirely locally with no external API calls, storing data in a user-specified directory with SQLite for reliable persistence and vector indices for efficient similarity search, making it valuable for AI assistants that need to remember user preferences, project configurations, coding patterns, and other contextual information across conversations while maintaining complete privacy and data control.