Agentic RAG with MCP Server is a powerful project that brings together an MCP (Model Context Protocol) server and client for building Agentic RAG (Retrieval-Augmented Generation) applications.
This setup empowers your RAG system with advanced tools such as:
- 🕵️♂️ Entity Extraction
- 🔍 Query Refinement
- ✅ Relevance Checking
The server hosts these intelligent tools, while the client shows how to seamlessly connect and utilize them.
Powered by the FastMCP class from the mcp library, the server exposes these handy tools:
| Tool Name | Description | Icon |
|---|---|---|
get_time_with_prefix |
Returns the current date & time | ⏰ |
extract_entities_tool |
Uses OpenAI to extract entities from a query — enhancing document retrieval relevance | 🧠 |
refine_query_tool |
Improves the quality of user queries with OpenAI-powered refinement | ✨ |
check_relevance |
Filters out irrelevant content by checking chunk relevance with an LLM | ✅ |
The client demonstrates how to connect and interact with the MCP server:
- Establish a connection with
ClientSessionfrom themcplibrary - List all available server tools
- Call any tool with custom arguments
- Process queries leveraging OpenAI or Gemini and MCP tools in tandem
- Python 3.9 or higher
openaiPython packagemcplibrarypython-dotenvfor environment variable management
# Step 1: Clone the repository
git clone https://github.com/ashishpatel26/Agentic-RAG-with-MCP-Server.git
# Step 2: Navigate into the project directory
cd Agentic-RAG-with-MCP-Serve
# Step 3: Install dependencies
pip install -r requirements.txt- Create a
.envfile (use.env.sampleas a template) - Set your OpenAI model in
.env:
OPENAI_MODEL_NAME="your-model-name-here"
GEMINI_API_KEY="your-model-name-here"- Start the MCP server:
python server.py- Run the MCP client:
python mcp-client.pyThis project is licensed under the MIT License.
Thanks for Reading 🙏