GET https://api.weather.gov/points/{latitude},{longitude} using WGS84 decimal coordinates to retrieve the metadata for that location, including the 'forecast' and 'forecastHourly' URLs specific to the grid cell
Follow the 'forecastHourly' URL from the previous response to GET the hourly forecast JSON for the next 7 days, which includes temperature, wind speed, precipitation probability, and weather description per hour
Set an 'Accept: application/geo+json' or 'Accept: application/ld+json' header to receive structured data; requests without an Accept header may return unexpected formats
Cache the grid point metadata response (step 1) by coordinate since the gridpoint assignment rarely changes; only re-fetch the forecast URL (step 2) to get fresh weather data
Parse 'properties.periods' from the forecast response, extracting 'startTime', 'temperature', 'windSpeed', 'shortForecast', and 'probabilityOfPrecipitation' for each period
Known gotchas
The NWS API only covers US locations including territories; requests for coordinates outside the US return a 404 — for international weather data a different provider is required
The API has no authentication requirement but enforces a User-Agent policy; requests without a descriptive User-Agent header (contact email or app name recommended) may be rate-limited or blocked
The forecast grid data endpoint occasionally returns a 500 error during high-demand periods (major weather events); implement retry logic and consider caching the last successful response as a fallback
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp