Open-Meteo API

Open-Meteo API

Open-Meteo is a free, open-source weather API. No API key, no signup for non-commercial use (fair-use limit ~10,000 requests/day). It blends national weather models (NOAA GFS, DWD ICON, ECMWF, Météo-France, JMA...) and auto-selects the best model per location.

Fastest path: bundled script

For the common case — current conditions and a daily forecast for a place — run the bundled zero-dependency script (Python 3.8+, stdlib only) instead of composing API calls by hand:

python3 scripts/get_weather.py "Taipei"                 # current + 3-day forecast
python3 scripts/get_weather.py "Paris" --country FR     # disambiguate place name
python3 scripts/get_weather.py --lat 25.05 --lon 121.53 # skip geocoding
python3 scripts/get_weather.py "Tokyo" --days 7 --json  # raw JSON for parsing
python3 scripts/get_weather.py "Denver" --fahrenheit

It geocodes, fetches, and decodes WMO weather codes into readable text. Exit codes: 0 success, 1 place not found, 2 API/network error.

Call the API directly (below) for anything the script doesn't cover: hourly data, specific variables (UV, soil, radiation...), historical weather, air quality, marine, or non-Python environments.

Typical workflow (direct API)

Most user questions name a place, but the API takes coordinates:

1. Geocode the place name → latitude/longitude (Geocoding API below). Skip this step if coordinates are already known.
2. Fetch weather from the forecast (or other) endpoint.
3. Interpret weather_code using the WMO table below.

Quick reference

All endpoints: GET with query parameters, JSON responses, HTTPS, CORS enabled. See references/other-apis.md for historical/air-quality/marine/elevation/flood/climate details.

Forecast API

GET https://api.open-meteo.com/v1/forecast

Key parameters:

Most-used variables (full list in references/weather-variables.md):

• current: temperature_2m, relative_humidity_2m, apparent_temperature, precipitation, weather_code, wind_speed_10m, wind_direction_10m, is_day
• hourly: temperature_2m, precipitation_probability, precipitation, weather_code, wind_speed_10m, relative_humidity_2m, cloud_cover, uv_index
• daily: weather_code, temperature_2m_max, temperature_2m_min, precipitation_sum, precipitation_probability_max, sunrise, sunset, uv_index_max, wind_speed_10m_max

Example: current weather

GET https://api.open-meteo.com/v1/forecast?latitude=25.05&longitude=121.53
    &current=temperature_2m,relative_humidity_2m,apparent_temperature,precipitation,weather_code,wind_speed_10m
    &timezone=auto

{
  "latitude": 25.06, "longitude": 121.52,
  "timezone": "Asia/Taipei", "utc_offset_seconds": 28800, "elevation": 12.0,
  "current_units": {"temperature_2m": "°C", "wind_speed_10m": "km/h", "weather_code": "wmo code"},
  "current": {
    "time": "2026-07-03T15:00", "interval": 900,
    "temperature_2m": 31.8, "relative_humidity_2m": 63,
    "apparent_temperature": 37.6, "precipitation": 0.0,
    "weather_code": 1, "wind_speed_10m": 5.1
  }
}

Example: 3-day daily forecast

GET https://api.open-meteo.com/v1/forecast?latitude=25.05&longitude=121.53
    &daily=weather_code,temperature_2m_max,temperature_2m_min,precipitation_sum,precipitation_probability_max,sunrise,sunset
    &timezone=auto&forecast_days=3

{
  "daily_units": {"temperature_2m_max": "°C", "precipitation_sum": "mm"},
  "daily": {
    "time": ["2026-07-03", "2026-07-04", "2026-07-05"],
    "weather_code": [95, 95, 96],
    "temperature_2m_max": [33.6, 33.5, 33.2],
    "temperature_2m_min": [26.0, 25.7, 26.1],
    "precipitation_sum": [1.5, 4.3, 4.2],
    "precipitation_probability_max": [100, 99, 82],
    "sunrise": ["2026-07-03T05:08", "..."],
    "sunset": ["2026-07-03T18:47", "..."]
  }
}

Response shape: hourly/daily data comes as parallel arrays — time[i] pairs with temperature_2m_max[i]. Units for every requested variable are echoed in *_units. Timestamps are ISO 8601 in the requested timezone (local time when timezone=auto), without a UTC offset suffix.

Geocoding API

GET https://geocoding-api.open-meteo.com/v1/search?name=Taipei&count=5&language=en&format=json

Parameters: name (place name, fuzzy for ≥3 chars), count (1–100, default 10), language (lowercase code: en, zh, ja, de...).

{
  "results": [
    {
      "id": 1668341, "name": "Taipei",
      "latitude": 25.05306, "longitude": 121.52639,
      "country": "Taiwan", "country_code": "TW",
      "admin1": "Taiwan", "timezone": "Asia/Taipei", "population": 7871900
    }
  ]
}

Pick the first result unless the user's context suggests otherwise (check country/admin1 when names are ambiguous, e.g. Paris TX vs Paris FR; results are ordered by relevance and population). No results key at all means no match — retry with a simpler query (city name only, no district/street; the geocoder matches localities, not addresses).

WMO weather codes

weather_code values (WMO 4677 subset) — condensed table:

Code examples

import requests

def get_weather(place: str) -> dict:
    geo = requests.get(
        "https://geocoding-api.open-meteo.com/v1/search",
        params={"name": place, "count": 1}, timeout=10,
    ).json()
    if "results" not in geo:
        raise ValueError(f"Place not found: {place}")
    loc = geo["results"][0]
    r = requests.get(
        "https://api.open-meteo.com/v1/forecast",
        params={
            "latitude": loc["latitude"], "longitude": loc["longitude"],
            "current": "temperature_2m,weather_code,wind_speed_10m",
            "daily": "temperature_2m_max,temperature_2m_min,precipitation_probability_max",
            "timezone": "auto",
        }, timeout=10,
    )
    r.raise_for_status()
    return r.json()

const geo = await (await fetch(
  `https://geocoding-api.open-meteo.com/v1/search?name=${encodeURIComponent(place)}&count=1`
)).json();
const { latitude, longitude } = geo.results[0];
const weather = await (await fetch(
  `https://api.open-meteo.com/v1/forecast?latitude=${latitude}&longitude=${longitude}` +
  `&current=temperature_2m,weather_code&timezone=auto`
)).json();

Errors

Errors return HTTP 400 with {"error": true, "reason": ""} — e.g. missing longitude, an unknown variable name, or an out-of-range date. Read reason; it states exactly which parameter is wrong.

Practical tips

• Request only the variables you need — response size and quota cost scale with variable count and day range.
• Always pass timezone=auto when using daily (required) or showing times to a user; otherwise times are GMT and daily aggregation boundaries are wrong.
• "Will it rain?" → prefer precipitation_probability (hourly) or precipitation_probability_max (daily) over raw precipitation amounts.
• Yesterday/last week → use past_days on the forecast endpoint. Older than ~3 months → use the archive endpoint (see references/other-apis.md). Archive data lags real time by ~5 days.
• The returned latitude/longitude are the model grid cell center (~1–11 km grid), so they differ slightly from the request — that's normal.
• interval in current is the data refresh interval in seconds; current conditions are model data, not station readings.
• Batch multiple locations in one call: latitude=25.05,24.14&longitude=121.53,120.68 returns a JSON array of result objects.
• Free tier is non-commercial use only; commercial use needs a paid API key (&apikey= on customer- prefixed endpoints). No key is needed otherwise.

For the complete variable catalog read references/weather-variables.md; for historical, air-quality, marine, elevation, flood, climate, and ensemble endpoints read references/other-apis.md.