Weather API

Base URL: https://api.wxalerts.org

One endpoint. Give it a point, and it returns what is happening there now and what is expected, sized for something that renders it.

GET /weather?lat=33.215&lon=-97.133
GET /weather?q=denton+tx
GET /weather?location_id=12345

No key needed. Browsers block calls from other websites (CORS allows only wxalerts.org), so call it from a server, script or app.

{
  "server_time": "2026-11-09T19:13:00Z",
  "location": {
    "name": "Denton", "display_name": "Denton, TX", "state": "TX",
    "lat": 33.2174, "lon": -97.1413, "timezone": "America/Chicago"
  },
  "units": { "system": "us", "fields": { "temperature": { "unit": "F", "label": "°F" } } },
  "current": {
    "observed_at": "2026-11-09T19:11:00Z",
    "sky": "Mostly cloudy", "icon": "cloudy", "is_day": true,
    "temperature": 68.0, "dewpoint": 57.9, "relative_humidity": 70.0,
    "wind_speed": 1.0, "wind_gust": 1.0,
    "wind_direction": 179.0, "wind_direction_cardinal": "S",
    "pressure": 30.21, "visibility": 10.0,
    "precip_1h": 0.0, "precip_24h": 0.0, "snow_depth": null,
    "conditions": null,
    "station": {
      "station_id": "aprs:KC5CAM-13", "name": null,
      "source": "aprs", "distance_m": 2094, "age_seconds": 121
    }
  },
  "hourly": [
    {
      "time": "2026-11-09T13:00:00-06:00",
      "sky": "Mostly cloudy", "icon": "cloudy", "is_day": true, "precip_type": "rain",
      "temperature": 70.6, "dewpoint": 59.8, "relative_humidity": 68.9,
      "wind_speed": 4.3, "wind_gust": 8.7,
      "wind_direction_cardinal": "SSW",
      "cloud_cover": 83.1, "precip_probability": 3.0, "precip_amount": 0.0
    }
  ],
  "daily": [
    {
      "date": "2026-11-09", "hours": 11,
      "sky": "Mostly cloudy", "icon": "cloudy", "precip_type": "rain",
      "high": 71.3, "low": 61.1, "dewpoint_high": 61.6,
      "wind_speed": 5.4, "wind_gust": 10.1, "wind_direction_cardinal": "SSW",
      "cloud_cover": 73.7, "precip_probability": 12.0, "precip_amount": 0.005
    }
  ]
}

Now is measured. The rest is forecast.

current comes from the surface weather stations around the point: airfield observations, mesonets and CWOP stations. It is not the model’s opinion of the present. A model’s analysis hour is still a forecast however recently it ran, and there is usually a thermometer within a few miles that is not guessing.

hourly and daily are the National Blend of Models. The two are never blended together and they can disagree, which is not a bug: an observation is a fact about this minute and a forecast is a guess about the next one.

Both start where the reader is standing in time: hourly[0] is the hour containing server_time and daily[0] is today at the forecast location. A cycle’s series really begins in the past (NBM runs every hour and its first step is an hour after the cycle behind it) and that reach-back is cut off here so that a strip can be drawn straight from the array. GET /forecast serves the same series whole, past hours included, for clients charting a run rather than displaying it.

station names the instrument the reading is attributed to: the one that supplied most of the fields. Where it has no sensor for something, the next station nearby fills that field silently. In the example the backyard station has no visibility sensor, so the 10.0 came from the airfield 6 km away. name comes from the station roster: airfields are named, CWOP stations usually are not, and it is null there.

current.sky and current.icon are the exception to “measured”. No station reports total cloud cover, so they come from the forecast hour nearest server_time. is_day is not borrowed: it is the sun’s position at this instant and this point.

current is null when no station within 50 miles has reported in the last two hours. That is a different statement from every field coming back empty, and it is deliberate: a reading from an airfield 200 miles away is a true reading and a false claim about here.

conditions is the station’s own words for what is falling, in METAR code: "-RA BR", "HZ". It is null when the station reports none, which is not the same as fair weather. For a phrase covering the sky as well, read hourly[0].sky.

One word for the sky

Every hour and every day carries sky: a single phrase, decided on the server.

precip_type is one of rain, snow, sleet or freezing rain: the likeliest kind should anything fall. It is set whenever any type has a chance above zero, so read it beside precip_probability (the example’s hour says rain at a 3% chance). It is null when every type is zero.

One picture

icon is the same decision as sky, made as a picture: clear-day, clear-night, partly-cloudy-day, partly-cloudy-night, cloudy, overcast, rain, snow, sleet, freezing-rain or thunderstorm. A client supplies artwork for those eleven and nothing else. At a 25% chance of precipitation or more it draws the precipitation (thunderstorm with a thunder probability ≥ 30%); below that it draws the cloud.

Day and night come from is_day, which is the sun’s position at that hour and place rather than the clock, because “day is 6 to 6” is wrong for months at a time in Alaska. Days carry no is_day and their icon is always the daytime one: a card for Thursday showing a moon would say Thursday is night.

The store publishes numbers, not prose, so this wording is derived. It is derived here rather than in each client on purpose: the underlying inputs are four independent precipitation-type probabilities that do not sum to 100, a cloud fraction and a thunder probability, and two panels in one house deriving their own answers is how they end up disagreeing about the same hour.

Fields

A null almost always means nothing publishes that field at that location, not that something failed.

current

Measured. Every one of these keys is always present; the value is null where no station nearby reports it.

hourly

From the hour containing server_time; see above.

daily

Local calendar days at the forecast location, not UTC days, from today onward. A UTC boundary falls in the evening across the Americas, which files an evening thunderstorm under tomorrow and puts tonight’s low on the wrong card.

How much to ask for

GET /weather?lat=33.2&lon=-97.1&hours=12&days=5

A panel showing the next twelve hours should ask for twelve. hours counts steps, not clock hours: the series is hourly for about two days, 3-hourly to about eight, then 6-hourly to about eleven, so hours=168 returns all of it, around a hundred steps. These only decide how much of the forecast crosses the wire, not how far ahead it reaches.

Units

GET /weather?lat=33.2&lon=-97.1&units=us

Every response carries a units block naming the unit of every field it returned, so a client never has to infer whether pressure came back in hPa or inHg. A gauge with a mislabeled scale looks entirely correct.

Observed and forecast fields share one vocabulary and one map. A temperature in current and a temperature in hourly are the same kind of number in the same unit, so both render through the same code, including high and low, which are declared as temperatures rather than leaving you to work it out.

Coverage

The forecast covers the United States and its territories (the Lower 48, Alaska, Hawaii, Puerto Rico and Guam) and nowhere else. A point outside it returns 200 with hourly and daily empty, not 404. There is no error to handle; there is simply no model there.

current is not bounded that way. CWOP and the airfield network are worldwide, so a point with no forecast can still have a thermometer next to it: London gets a current and an empty forecast.

Polling

See rate limits. The response is cached for 60 seconds, which is the station’s cadence rather than the model’s: CWOP stations report every five minutes and METARs hourly, while NBM runs once an hour. Polling faster than a minute gains nothing; polling once a minute is the most a display can usefully do.