Query daily weather forecast from Open-Meteo using coordinates or place names with geocoding support. Use when: weather forecast, upcoming weather, multi-day weather outlook.
This skill fetches weather forecast data from Open-Meteo and returns a normalized structure suitable for LLM parsing. You SHOULD set only the fields necessary for your use case to avoid hitting response length limits. Prefer fetching fewer forecast days or hours. You CAN rephrase the returned data in a more natural way or extract specific details as needed.
Call the run_js tool using index.html and a JSON string for data with the following fields:
location: Optional string. Preferred when coordinates are not known. Required if latitude/longitude are not provided. Used for geocoding (for example, "California", "White House", or "here" to auto-derive from IP). Use the smallest known location area in this field.location_filter: Optional string or array. Extra geocoding filter applied client side. Use this for broader qualifiers that should not be part of location (for example, "US", "United States", "Missouri").country_code: Optional string. Two-letter country filter for geocoding (for example, )."US"result_index: Optional integer. Choose a specific geocoding candidate when the location is ambiguous.geocode_count: Optional integer from 1 to 20. Number of geocoding candidates to fetch. Defaults to 10.latitude: Optional number. Must be paired with longitude.longitude: Optional number. Must be paired with latitude.timezone: Optional string. IANA timezone or GMT offset (for example, Asia/Tokyo, GMT+8) or auto. Defaults to auto. Note: auto is buggy; supply an explicit timezone or offset when possible, derive it from another tool, or guess from context.forecast_mode: Optional string. day or hour. Defaults to day.forecast_days: Optional integer from 1 to 16. Defaults to 3. Only used in day mode.forecast_hours: Optional integer from 1 to 24. Defaults to 6. Only used in hour mode. DO NOT set this higher than necessary to avoid hitting response length limits.forecast_fields: Optional string or array. Controls what forecast fields are returned. Accepts comma-separated values or a list. Use all to return every available field for that mode. Leave empty for default set (recommended). DO NOT request more fields or all unless absolutely necessary. See the Forecast Field Options below for available fields.Use non-cryptic field names below for forecast_fields.
Day mode fields (forecast_mode: day):
date: Forecast date (also always included in each day entry).weather: Human-readable weather text.weather_code: WMO weather code.temperature: Object with both max and min temperatures.feels_like: Object with both max and min apparent temperatures.precipitation: Object including total precipitation, precipitation probability max, precipitation hours, and rain/showers/snowfall totals.sun: Object including sunrise, sunset, daylight duration, and sunshine duration.uv: Object including max UV and clear-sky max UV.wind: Object including max wind speed, max gusts, and dominant direction.Day mode default fields:
weathertemperatureprecipitationwindHour mode fields (forecast_mode: hour):
time: Forecast time (also always included in each hour entry).weather: Human-readable weather text.weather_code: WMO weather code.temperature: Hourly temperature.feels_like: Hourly apparent temperature.humidity: Hourly relative humidity.precipitation: Object including precipitation total, precipitation probability, rain, showers, and snowfall.snow_depth: Hourly snow depth.dew_point: Hourly dew point.visibility: Hourly visibility.surface_pressure: Hourly surface pressure.Hour mode default fields:
weathertemperaturehumidityprecipitationlocation to the special value here — the skill will derive coordinates and a best-guess timezone from the caller IP and include derived fields (location.latitude, location.longitude, location.city, location.region, location.country). If multiple timezones are returned by the lookup, the skill will guess the first one.location; the script resolves coordinates automatically via Open-Meteo geocoding."california, US"), put only the narrow place in location (for example, "california") and pass the broader qualifier in location_filter (for example, "US").country_code and/or result_index.timezone values whenever possible; use auto only as a fallback.geocoding summary (query, selected index/name, and candidate count) when geocoding is used.