Open-Meteo MCP. Predict every variable—from pollution to wind speed.

Listen up. This Open-Meteo MCP Server lets you pull global weather and air quality data straight into your agent without needing a single API key. You just connect your AI client, and boom—you got the raw environmental metrics you need.

You'll start by finding coordinates using get_geocoding. Just give it a readable place name, and that tool spits out the precise latitude and longitude needed for everything else.

From there, if you wanna know what the area's altitude is, run get_elevation with those coordinates; it returns the reading in meters. If you gotta pull historical context, use get_historical_weather. This tool pulls detailed weather logs—we're talking decades back to 1940—for any specific coordinate range and date.

When you’re forecasting what's coming up, get_forecast provides current conditions and predicts future metrics like temperature, wind direction, and rain probability. It gives you predictions for the next sixteen days, covering everything from basic weather to cloud cover.

To check out air quality—and this is key—you use get_air_quality. This tool calculates hourly pollution levels, including PM2.5 and Ozone readings, for up to seven days at specific coordinates. You can cross-reference that with the UV index predictions when you run it.

It’s all about location data first. Your agent runs get_geocoding on a city name, gets the lat/long pair, then passes those numbers to get_forecast for the next two weeks of weather predictions, or maybe to get_air_quality if you're checking pollution levels over seven days. If you need altitude context, it’s another shot at get_elevation.

For deep dives into climate trends, you feed coordinates and date ranges into get_historical_weather. The whole system works together like clockwork.

You don't have to worry about rates, keys, or complex authentication. You just tell your agent what job it needs doing—whether that’s predicting the wind speed for tomorrow at a specific coordinate pair using get_forecast, determining the historical rain chances from 1985 using get_historical_weather, or figuring out if the ozone levels are spiking next week by running get_air_quality.

Think of it this way: you've got five powerful tools. You start with get_geocoding to lock down your target spot. Then, you branch out based on what data you need. If it’s a short-term prediction, hit up get_forecast. If you're tracking pollution over time, use get_air_quality. If the job is pure history, get_historical_weather has your back.

And if you just gotta know how high up that spot is, get_elevation gives it to ya in meters. You can stack these calls together using your AI client; for example, running get_geocoding, then passing the coordinates to both get_forecast and get_air_quality so you get a full picture—the weather prediction alongside the pollution forecast.

This open-source approach means your agent runs complex queries, pulling data from modern predictions all the way back through archived records dating back decades. You never hit a wall because of required credentials or restricted variables.