jaredevans/mcp-rainsoon

Rainsoon MCP

This project provides a simple MCP (Model Context Protocol) server for the Gemini CLI that exposes a single tool: check_for_rain and also a /rainsoon slash command for quick access.

---

What is MCP?

MCP (Model Context Protocol) allows the Gemini CLI to securely and reliably interact with local or remote tools.

The rainsoon_mcp.py script runs as a local MCP server, making its check_for_rain function available as a callable tool inside Gemini CLI.

With the addition of an MCP prompt, Gemini CLI now exposes /rainsoon as a native slash command with arguments.

---

Features

check_for_rain(ip: str = "", threshold: int = 20) -> dict

• Checks the probability of rain for a specified IP address.
• If no IP address is provided, it automatically detects your public IP using multiple fallback services.
• You can set a custom threshold (in percent) to define what “rain soon” means (default: 20%).
• Returns a JSON object containing:
  • Location details
  • Rain probability
  • Whether the probability exceeds the threshold
  • A human-readable message

---

/rainsoon Command in Gemini CLI

The MCP server includes a prompt definition that Gemini CLI turns into a slash command:

Usage:

/rainsoon                      # auto-detect IP, default threshold=20%
/rainsoon --ip="134.231.2.45"       # check for a specific IP
/rainsoon --ip="134.231.2.45" --threshold=35

The output is the raw JSON from check_for_rain.

Example:

╭───────────────╮
│  > /rainsoon  │
╰───────────────╯

✔  check_for_rain (rainsoon MCP Server) {"threshold":20,"ip":""}

{
  "ip": "96.231.151.145",
  "location": "Wheaton",
  "lat": 39.0358,
  "lng": -77.0523,
  "rain": false,
  "precipitation_chance": 0,
  "threshold": 20,
  "message": "No, 0% chance of rain soon in Wheaton.",
  "hour_sample": [
    "2025-08-12T00:00",
    "2025-08-12T01:00",
    "2025-08-12T02:00"
  ]
}

---

How it Works

The rainsoon_mcp.py script is a self-contained MCP server. Its workflow:

1. IP Address Discovery

   • If ip is provided, uses it.
   • If not, queries a series of external services (api.ipify.org, ipinfo.io/ip, icanhazip.com) to detect your public IP.

2. Geolocation

   • Uses the geocoder library to get latitude, longitude, and city from the IP address.
   • If geolocation fails for a provided IP, it will attempt to fall back to your detected public IP.

3. Weather Data

   • Queries the Open-Meteo API for hourly precipitation probability for the next 24 hours at the coordinates.

4. Result

   • Compares the precipitation probability for the next hour to the threshold.
   • Returns detailed JSON with the location, chance of rain, and a boolean flag.

---

Installation

Install Python packages:

pip install "mcp[cli]" requests geocoder

Gemini CLI configuration (~/.gemini/settings.json):

{
  "selectedAuthType": "gemini-api-key",
  "theme": "Default",
  "mcpServers": {
    "rainsoon": {
      "command": "/absolute/path/to/venv/bin/python3",
      "args": ["-u", "/absolute/path/to/rainsoon_mcp.py"],
      "timeout": 15000
    }
  }
}

Fields:

• mcpServers — all MCP server configs.
• rainsoon — server name (used in /mcp call commands).
• command — absolute path to Python in your venv.
• args — -u for unbuffered output, then absolute path to the script.
• timeout — in ms.

⚠ Update the paths for your environment.

---

How to Use

Natural Language

Gemini CLI can auto-call the tool when you type:

Will it rain soon?
Will it rain soon for IP 134.231.2.45?
Will I be wet soon?

/rainsoon Command

/rainsoon
/rainsoon --ip="134.231.2.45"
/rainsoon --ip="134.231.2.45" --threshold=35

---