PlayMCP

You are managing and querying reservations across multiple platforms from a single MCP server. This setup lets you view bookings, prevent duplicates, handle revenue and customer management, and generate reminder messages across platforms like Naver, Airbnb, SpaceCloud, Yanolja, and Kakao.

How to use

Run the MCP server in either STDIO (local) or HTTP (remote) mode, then interact with it from an MCP client. The server exposes tools to retrieve today’s reservations, search by date or platform, check for duplicate bookings, manage revenue summaries, handle blacklists, and generate reminder messages. You can also synchronize with iCal sources to keep reservations up to date.

How to install

Prerequisites: Python 3.12 or newer must be installed on your system.

Create and activate a virtual environment, then install dependencies.

# 가상환경 생성
python3 -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate

# 의존성 설치
pip install -r requirements.txt

Test data and MCP server tests help verify your setup before going into production.

# 가상환경 활성화
source venv/bin/activate

# 샘플 데이터 생성
python test_data.py

Run the MCP server in STDIO mode for local use (Claude Desktop CLI).

python3 -m src.server

If you prefer HTTP transport for a remote MCP server, start with HTTP mode and set the transport accordingly. For example, you can start the server with HTTP transport on port 8000.

# HTTP transport로 실행
TRANSPORT=http PORT=8000 python3 -m src.server

Additional sections

Claude Desktop integration is supported via a configuration file that points to the MCP server command. You can copy a configuration snippet into Claude’s config directory to connect to the server and enable the tools described in this guide.

If you plan to run in a container or deploy to a remote environment, you can use the same commands you run locally, but ensure environment variables and working directories are set correctly for the deployment target.

Note on environment variables: you may need to propagate PATH or other variables to ensure the Python virtual environment and dependencies are visible to the MCP server process.

Available tools

get_today_reservations

Fetches reservations for today from all integrated platforms.

get_reservations_by_date

Retrieves reservations for a specific date across platforms.

get_reservations_by_platform

Aggregates reservations by platform (Naver, Airbnb, SpaceCloud, Yanolja, Kakao).

get_week_reservations

Provides reservations for the current week.

check_duplicate_reservations

Detects potential duplicate bookings across platforms.

check_time_slot

Verifies availability for a given time slot.

get_revenue_summary

Summarizes revenue by day, week, or month.

calculate_platform_fees

Calculates platform commissions.

get_blacklist

Lists blacklisted customers.

add_to_blacklist

Adds a customer to the blacklist.

get_noshow_customers

Retrieves customers with no-show history.

generate_reminder_message

Generates automatic reservation reminder messages.

sync_platform_ical

Synchronizes reservations from iCal sources.