KnowAir Weather MCP Server is a modular MCP implementation that sources data from Caiyun Weather API to deliver real-time weather, air quality, forecasts, and astronomical information via the Model Context Protocol. In the v2.0 architecture, the codebase is decomposed into dedicated modules: utils.py for data processing, config.py for centralized configuration, models.py for Pydantic validation, and server.py containing the MCP server core logic, enabling easier maintenance and extension. The server exposes a comprehensive API surface, including: real-time data (temperature, humidity, wind, visibility, pressure), minute-level precipitation forecasts, hourly forecasts up to 360 hours, daily forecasts up to 15 days, and historical weather data for the past 72 hours; extended ranges support 1-360 hours and 1-15 days. It also provides weather alerts, an integrated comprehensive interface get_comprehensive_weather, and astronomy data (sunrise/sunset, moon phase, moonrise/moonset). The platform emphasizes Chinese localization, intelligent formatting (emojis, friendly displays), and a unified API management layer that centralizes all Caiyun API calls. It is designed for developers integrating weather and air quality into MCP-enabled assistants and applications, with easy setup via environment variables and a quick start.