Price Monitor MCP Server

outline

This project is a price monitoring server based on Model Context Protocol (MCP). It compares the DB standard price and Gmarket real-time price using product codes, and sends a notification to Slack when the price drops.

Server/tool/prompt structure following MCP standards
Automate the entire process of crawling, price comparison, and notifications
Supports Slack webhook integration

Key Features

DB standard price search : Search the standard price in DB using the product code
Gmarket Real-time Price Crawling : Using Firecrawl API
Compare prices and calculate discount rates
Send Slack notifications when price drops
Provides a tool to automate the entire workflow

Folder structure

price_monitor_mcp/
├── src/
│   └── price_monitor_mcp.py   # MCP 서버 메인 코드
├── mcp_client.py              # MCP 클라이언트 코드
├── README.md                  
└── .env                       # 환경변수

How to run

1. Prepare virtual environment and install package

conda activate price_monitor_mcp
pip install -r requirements.txt
# 또는 필요한 경우
pip install mcp firecrawl requests python-dotenv psycopg2-binary pydantic

2. Setting environment variables

Set environment variables such as Slack webhooks in .env file as follows.

SLACK_WEBHOOK_URL=https://hooks.slack.com/services/...
DB_HOST=...
DB_PORT=...
DB_NAME=...
DB_USER=...
DB_PASSWORD=...

3. Run the MCP server

mcp run src/price_monitor_mcp.py

Or run it with devtools: mcp dev src/price_monitor_mcp.py
Or activate the conda environment with a shell script and then run

MCP Tools/Prompts List

get_db_price(product_code) : DB standard price query
crawl_gmarket_price(product_code) : Gmarket real-time price crawling
send_slack_alert(message) : Send Slack notification
monitor_price_workflow(product_code) : Automatically run the entire process (recommended)
monitor_price(product_code) : prompt (for LLM)

Automate the entire process (recommended)

Example of calling a workflow tool

In the MCP dev tools/client:

Select the monitor_price_workflow tool, enter product_code , and run it.
Results: Returns DB price, lowest price, price difference, discount rate, Slack notification, etc.

Python client example

import asyncio
from mcp import ClientSession, StdioServerParameters
from mcp.client.stdio import stdio_client

async def main():
    server_params = StdioServerParameters(
        command="python",
        args=["src/price_monitor_mcp.py"],
    )
    async with stdio_client(server_params) as (read, write):
        async with ClientSession(read, write) as session:
            await session.initialize()
            result = await session.call_tool("monitor_price_workflow", arguments={"product_code": "ULCK25151"})
            print("워크플로우 결과:", result)

if __name__ == "__main__":
    asyncio.run(main())

Slack notification test

If your Slack webhook is set up correctly, you will automatically be notified when a price drops.
The message format can be freely modified in send_slack_alert function.

LLM (Claude, GPT, etc.) linkage

Claude, GPT, etc. will be able to run prompts/tools in natural language once MCP server connection is officially supported
Currently, the results are received through MCP client code and pasted to LLM for summary/analysis request.

References/Documents

MCP Python SDK Official Documentation
Firecrawl, Slack API, DB, etc. need to be set up for each environment.

Contact/Contribution

Please leave any questions, bugs, or extension requests as issues!

Integrations