Find the places a name could refer to, with coordinates and timezone.

Use this when a place name is ambiguous and the context does not settle it -- 'Springfield' is five US cities, 'Portland' two large ones - or when you want coordinates to pass to the other tools directly.

The other tools accept a place name themselves and do not need this first. Reach for it to disambiguate, not as a routine preliminary step.

Daily weather forecast for a place: highs, lows, rain and wind, day by day.

The main tool. Answers "what is the weather doing this week", "how warm will it be in Boston next Tuesday", and "will it rain while I am there".

Give dates when you have them -- from a calendar event, say -- and they are read as local dates at the destination, not wherever this server is running. Give neither and it covers the next week starting from today at that location, which is the right anchor when the place is in another timezone.

Forecasts reach at most 16 days ahead; that is a limit of the data. For a trip further out, use get_typical_weather, which reports what the weather actually did on those dates in past years rather than pretending to predict.

Hour-by-hour forecast for one day: when exactly the rain arrives.

Use this once the day matters more than the week -- "will it rain during the 2pm walk", "is the morning or the afternoon better for the drive". For a multi-day outlook use get_forecast instead; this returns 24 rows and is wasteful for a question about a week.

The day must be within the 16-day forecast horizon, and is read as a local date at the location.

What the weather actually did on these dates in past years. Not a forecast.

For trips beyond the 16-day forecast horizon, which is most trips at the point they get planned. Returns observed history: the average high and low for this window over recent years, the extremes, and how many of those years saw rain during it.

This is history and must be described as history. It says what to pack, not what will happen. Once the dates come within 16 days, get_forecast replaces it entirely — do not keep quoting averages when a real forecast exists.

The year is taken from the dates only to fix the calendar window; the summary always covers complete past years.

Rank the days ahead by how pleasant it will be outside. For planning a hike or similar.

Answers "when should I go", not "what is the weather". Use it when the user wants a day picked out of a range — a hike, a ride, a picnic, painting the fence — rather than the outlook for a day they have already chosen.

Days are scored over the hours actually spent outside (8am to 6pm by default), not over the whole day, because a daily high is one afternoon moment and a daily rain chance includes the hours you were asleep. The temperature scored is feels-like, which already accounts for humidity, wind chill and sun — so a muggy 85°F is correctly ranked below a dry one.

The ranking is a stated preference, not a measurement. Every day comes back with the numbers behind it — feels-like range, rain chance, wind, humidity — so report those alongside the pick rather than the score alone.