🌦️ Docs | Open-Meteo.com

Seamless integration of high-resolution weather models with up 16 days forecast
By default, we provide forecasts for 7 days, but you can access forecasts for up to 16 days. If you're interested in past weather data, you can use the Past Days feature to access archived forecasts.
Open-Meteo combines weather model output from multiple national weather services into a continuous, seamlessly updated timeseries. Each time a model run is ingested, the forecast data are stitched to the previous run without gaps or discontinuities, so the hourly timeseries always reflects the latest available initialisation. For each location, the highest-resolution applicable model is selected automatically.
Weather models cover different geographic areas at different resolutions and provide different weather variables. Depending on the model, data have been interpolated to hourly values or not all weather variables are available. Use the Weather models dropdown (just below the hourly variables) to select and compare individual models. To access the full archive of past forecast runs as issued, useful for forecast verification or training ML models, see the Historical Forecast API and the Single Runs API.
The default model selection best_match automatically combines the most suitable weather models for the requested coordinates. Every forecast is built on a global backbone of DWD ICON, NOAA GFS and ECMWF IFS. Depending on the location, high-resolution regional models are added on top. Regions are evaluated from top to bottom and the first match determines the model combination:
For each weather variable and forecast hour, the highest-resolution model takes precedence. Where it does not provide data — because its forecast length is exceeded, a variable is not available or the location lies outside its coverage — the next coarser model fills the gap with a smooth transition. Precipitation probability is always derived from ensemble models such as the GFS, ICON and ECMWF IFS ensembles. To override this selection, pick individual models in the Weather models section above.
The API endpoint /v1/forecast accepts a geographical coordinate, a list of weather variables and responds with a JSON hourly weather forecast for 7 days. Time always starts at 0:00 today and contains 168 hours. If &forecast_days=16 is set, up to 16 days of forecast can be returned. All URL parameters are listed below:
Additional optional URL parameters will be added. For API stability, no required parameters will be added in the future!
The parameter &hourly= accepts the following values. Most weather variables are given as an instantaneous value for the indicated hour. Some variables like precipitation are calculated from the preceding hour as an average or sum.
The parameter &minutely_15= can be used to get 15-minutely data. This data is based on NOAA RRFS model for North America and DWD ICON-D2 and Météo-France AROME model for Central Europe. If 15-minutely data is requested for other regions data is interpolated from 1-hourly to 15-minutely.
15-minutely data can be requested for other weather variables that are available for hourly data, but will use interpolation.
Pressure level variables do not have fixed altitudes. Altitude varies with atmospheric pressure. 1000 hPa is roughly between 60 and 160 meters above sea level. Estimated altitudes are given below. Altitudes are in meters above sea level (not above ground). For precise altitudes, geopotential_height can be used.
All pressure levels have valid times of the indicated hour (instant).
Aggregations are a simple 24 hour aggregation from hourly values. The parameter &daily= accepts the following values:
On success a JSON object will be returned.
In case an error occurs, for example a URL parameter is not correctly specified, a JSON error object is returned with a HTTP 400 status code.
(*) Codes 96 and 99 are only reported by models with an explicit hail forecast, such as DWD ICON or UKMO. All other models derive thunderstorms from instability parameters and report codes 95 and 97.
The API only returns numeric weather codes. To display descriptions, map the codes on the client. The mapping below can be copied directly.