Home / MCP / Haiguitang MCP Server

Haiguitang MCP Server

Hosts and serves Haiguitang turtle-soup puzzles via a local Python-based MCP server controlled by UV, enabling prompt retrieval and puzzle management.

python

7stars

View on GitHub

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
    "mcpServers": {
        "haiguitang_mcp": {
            "command": "uv",
            "args": [
                "--directory",
                "E:\\haiguitangmcp\\haiguitang_mcp",
                "run",
                "server.py"
            ]
        }
    }
}

You can run a local MCP server to host and interact with Haiguitang turtle-soup puzzles, letting you play solo or with clients that connect to the server. This setup uses a Python-based server controlled via the UV tool, enabling you to manage puzzles, prompts, and game rules from your workstation.

How to use

You will connect your MCP client to a local stdio MCP server that runs with the UV tool. Start the server locally, then your client can request puzzle data, prompts, and lists through the standard MCP interactions. You can fetch the full game rules template, pull a specific puzzle by title, or list all available puzzles. To begin, ensure you have the server running and then configure your MCP client to point at the local process.

How to install

Prerequisites you need before starting are Python installed on your computer and the UV tool available in your PATH.

Clone the project, install dependencies with UV, and prepare the local server.

git clone https://github.com/wangyafu/haiguitangmcp/
cd haiguitangmcp
uv sync

Additional configuration steps for MCP clients

You will configure your MCP client to launch the local Haiguitang MCP server via UV, using the specified directory and starting the server script.

"mcp":{
  "servers":{
    "haiguitang_mcp": {
      "type": "stdio",
      "command": "uv",
      "args": [
        "--directory",
        "E:\\haiguitangmcp\\haiguitang_mcp",
        "run",
        "server.py"
      ]
    }
  }
}

Additional configuration for Cherry Studio style setup

If you prefer configuring via Cherry Studio, use the following local MCP server entry to start the same Python server via UV.

"mcpServers": {
  
  "haiguitang": {
    "isActive": true,
    "name": "海龟汤MCP服务器",
    "description": "和用户玩海龟汤",
    "registryUrl": "",
    "command": "uv",
    "args": [
      "--directory",
      "E:/haiguitangmcp/haiguitang_mcp",
      "run",
      "server.py"
    ]
  },
   
}

Note on paths and variations

The path "E:/haiguitangmcp/haiguitang_mcp" indicates where server.py resides for the local MCP server. You can adapt the directory separator to your operating system as needed.

What you can do with this MCP server

You can fetch the full game rules template, obtain a specific puzzle by title, and list all available puzzles. You can also access puzzle information via the puzzle URL scheme puzzles://{puzzle_title} and use the game_rules template to guide play.

Troubleshooting tips

If the server does not start, verify that Python is installed and accessible, and that UV is installed and available on your system PATH. Confirm the directory paths you configured exist and contain server.py. Check for any typos in configuration blocks and ensure you launched the correct startup command.

Security and licensing notes

Ensure you respect any puzzle copyrights and attribute authorship where applicable when sharing or contributing puzzles. Follow standard security practices for local development servers and avoid exposing the server to untrusted networks.

Available tools

get_prompt

Retrieves the full prompt for Haiguitang turtle-soup game rules and setup.

get_puzzle

Fetches the complete content of a specific puzzle by title.

list_puzzles_tool

Lists all puzzles available in the MCP puzzle collection.

puzzles

URL-based access to puzzle information in the form puzzles://{puzzle_title}.

game_rules

Provides a template describing the standard rules and flow of the turtle-soup game.