LLM-MCP-RAG-JS

What it does

LLM-MCP-RAG-JS is a TypeScript-based MCP server that implements a complete RAG (Retrieval-Augmented Generation) pipeline. It allows an AI assistant to fetch specific knowledge from local documents, augment the prompt with relevant context, and generate accurate responses based on that data.

Tools

• load_documents: Loads and vectorizes documents from a specified directory.
• retrieve_context: Performs similarity searches to find the most relevant document snippets for a given query.
• process_query: Coordinates the retrieval and generation process to provide a context-aware answer.

Installation

Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "llm-mcp-rag-js": {
      "command": "node",
      "args": ["/path/to/openai-llm-mcp-rag-js/dist/index.js"],
      "env": {
        "OPENAI_API_KEY": "your_api_key",
        "OPENAI_BASE_URL": "https://api.openai.com/v1",
        "EMBEDDING_BASE_URL": "https://api.openai.com/v1",
        "EMBEDDING_KEY": "your_embedding_key"
      }
    }
  }
}

Supported hosts

• claude