Agentic-RAG-with-MCP-Server

Agentic-RAG-with-MCP-Server

3.4

If you are the rightful owner of Agentic-RAG-with-MCP-Server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

Agentic RAG with MCP Server is a project that integrates an MCP server and client to build Retrieval-Augmented Generation applications.

Agentic RAG with MCP Server is a comprehensive project designed to enhance Retrieval-Augmented Generation (RAG) systems by integrating a Model Context Protocol (MCP) server and client. This setup provides advanced tools for entity extraction, query refinement, and relevance checking, which are crucial for improving the accuracy and efficiency of RAG applications. The server hosts these intelligent tools, while the client demonstrates how to connect and utilize them effectively. By leveraging the power of OpenAI and other technologies, this project aims to streamline the process of retrieving and generating relevant information, making it a valuable asset for developers working on RAG systems.

Features

  • Entity Extraction: Uses OpenAI to extract entities from queries, enhancing document retrieval relevance.
  • Query Refinement: Improves the quality of user queries with OpenAI-powered refinement.
  • Relevance Checking: Filters out irrelevant content by checking chunk relevance with an LLM.
  • Current Date & Time: Provides the current date and time.
  • Seamless Integration: Demonstrates how to connect and interact with the MCP server using a client.

Tools

  1. get_time_with_prefix

    Returns the current date & time.

  2. extract_entities_tool

    Uses OpenAI to extract entities from a query.

  3. refine_query_tool

    Improves the quality of user queries with OpenAI-powered refinement.

  4. check_relevance

    Filters out irrelevant content by checking chunk relevance with an LLM.