Israel Weather Server

📌 Project Description

This project implements a custom MCP (Model Context Protocol) server that provides Large Language Models (LLMs) with real-time access to dynamic weather data in Israel.

Unlike traditional setups that rely on rigid static APIs, this project equips the model with Browser Automation capabilities using the Playwright library. The intelligent agent automatically launches a browser instance behind the scenes, navigates to a popular Israeli weather website (weather2day), searches for a specific city based on user input, and extracts the live weather data (RAG). This data is then fed back into the LLM's context window, enabling it to formulate highly accurate, real-time responses.

🎯 Project Goals & Architecture (Tools)

The MCP server exposes the following 4 tools to the LLM:

1. open_weather_forecast_israel: Launches a headless/headed Chromium browser and navigates to the target weather portal.
2. enter_weather_forecast_city_israel: Smart detection of the search field and user-like typing emulation (with delays) to smoothly bypass anti-bot mechanisms.
3. select_weather_forecast_city_israel: Interacts with the dynamic dropdown menu using keyboard emulation (Arrow keys + Enter) to load the city's specific weather page.
4. get_weather_forecast_data: Performs text extraction (Web Scraping / RAG Context feeding) from the body of the page, sanitizes the text, and returns a clean data chunk to the LLM.

🛠️ Technology Stack

• Python 3.10+
• MCP SDK (Anthropic's official Model Context Protocol implementation)
• Playwright (Microsoft's modern browser automation framework for Chromium)
• uv (An extremely fast Python package and environment manager)

🚀 Installation & Usage Instructions

1. Synchronize and install project dependencies:

   python -m uv sync