SuperNova RAG
Enables semantic search over internal documentation using a local server with Retrieval-Augmented Generation (RAG).
About
SuperNova RAG is a practical demonstration of building and running a local server that uses Retrieval-Augmented Generation (RAG) for semantic search over internal documentation. Built with Node.js and TypeScript, it leverages Hugging Face embeddings and an in-memory vector store for fast, context-aware answers. This allows developers to quickly find relevant information within their documentation, directly within tools like Cursor, enhancing productivity and reducing time spent searching for answers.
Key Features
- Hugging Face embeddings for semantic search
- 1 GitHub stars
- Model Context Protocol (MCP) server implementation
- In-memory vector store for fast retrieval
- Retrieval-Augmented Generation (RAG) pipeline
- Integration with Cursor and other MCP-compatible tools
Use Cases
- Semantic search over internal documentation
- Context-aware question answering for developers
- Integration with IDEs like Cursor for instant documentation access