🚀 Memory Data Server (MCP Tool Provider)

A lightweight Model Context Protocol (MCP) server built with Python and FastMCP. This service acts as a tool provider that enables LLMs and Agentic AI systems (e.g., LangChain, LangGraph, Claude Desktop) to inspect and manage customer support records via Streamable HTTP.

📌 Key Features

• Standardized MCP Interface: Exposes backend tools via the open Model Context Protocol.

• Streamable HTTP Transport: Operates over HTTP via MCP's current Streamable HTTP transport, making it Docker-friendly and easy to deploy in containerized environments.

• Deterministic Tool Capabilities: Exposes typed data-querying and mutation functions (fetch_customer_ticket, create_ticket, update_status) to provide dynamic context to LLM reasoning loops.

Related MCP server: CustomerOS MCP Server

🏗 System Architecture

┌─────────────────────────┐    Streamable HTTP (Port 8000)   ┌──────────────────────────────┐
│  AI Agent / LLM Host    │ ───────────────────────────────► │    FastMCP Tool Server       │
│  (LangChain / Claude)   │ ◄─────────────────────────────── │   (Memory Data Server)       │
└─────────────────────────┘                                  └──────────────┬───────────────┘
                                                                            │
                                                                            ▼
                                                                 ┌────────────────────┐
                                                                 │ In-Memory DB Lookup│
                                                                 └────────────────────┘

🛠 Tech Stack

• Language: Python 3.10+

• Framework: mcp (FastMCP)

• Transport Protocol: Streamable HTTP over HTTP

🚀 Getting Started

1. Prerequisites

Ensure you have Python installed, then install the dependencies:

pip install -r requirements.txt

2. Running the Server

Execute the script to start the FastMCP Streamable HTTP server bound to 0.0.0.0:8000:

python server.py

Note: The server will start and expose its MCP endpoint at http://0.0.0.0:8000/mcp.

🛠 Available Tools

fetch_customer_ticket

Performs a lookup for customer support tickets using a unique ticket ID.

• Input Parameters:

  • ticket_id (int, required): The numeric identifier of the support ticket.

• Output:

  • str: Formatted text string containing ticket owner, issue summary, customer tier, and status (or an error message if missing).

create_ticket

Creates a new customer support ticket with an auto-assigned ID and open status.

• Input Parameters:

  • user (str, required): Name of the customer opening the ticket.

  • issue (str, required): Description of the issue.

  • tier (str, optional, default "Free"): Customer tier.

• Output:

  • str: Formatted text string confirming the created ticket.

update_status

Updates the status of an existing ticket.

• Input Parameters:

  • ticket_id (int, required): The ticket to update.

  • status (str, required): One of open, in_progress, resolved, closed.

• Output:

  • str: Formatted text string confirming the update, or an error message if the ticket is missing or the status is invalid.

🧪 Quick Test (Verifying the MCP Endpoint)

While running the server, you can verify the Streamable HTTP endpoint is active with an MCP initialize request:

curl -i -X POST http://localhost:8000/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"curl-test","version":"1.0"}}}'

📜 Roadmap

• Basic FastMCP server with Streamable HTTP transport

• In-memory database simulation tool

• Implement write/mutation tools (create_ticket, update_status)

• Docker containerization setup

• Connect tool backend to persistent storage (PostgreSQL / AWS DynamoDB)

📄 License

This project is licensed under the MIT License — free to use, modify, and distribute.