sleevecast API guide

Overview

This API tells you what people wear outside in a city. It answers for the current hour and for each of the next days. Use it to show a hint such as "light jacket, closed shoes", or to help a traveller pack.

• What you get. For every populated place GeoNames lists with more than 15,000 people, every capital, and every city where we have cameras, the share of people who wear each value of five things: the upper layer, the lower body, footwear, a scarf and an open umbrella.
• Where it comes from. A weather model. Where we have street cameras in a city, what they see can correct the model.
• Where it lives. The base URL is https://api.sleevecast.app. This page is at https://api.sleevecast.app/guide/. The machine-readable description (OpenAPI 3.1) is at https://api.sleevecast.app/guide/openapi.json. It needs no key.
• Interactive reference. Read the API in your browser and try a call with your own key: /guide/reference/. It is served by this API and loads nothing from other sites.
• A typed client. Generate one from that document:

npx openapi-typescript https://api.sleevecast.app/guide/openapi.json

• The version. The contract is versioned. The latest version, and the date of the first frozen contract, are in the Changelog. GET /v1/version returns the same number as contract_version.

Every call to /v1 needs a key. The Quickstart shows your first three calls.

Quickstart

1. Get a key. Keys are issued by hand, to one organisation. Write to the address under Support. Read the conditions of use before you ask. A key expires after the number of days in the table under Limits and caching, unless we agree otherwise.
2. Find the city. Call nearest with a coordinate, and keep the slug of the city it returns. Round the coordinate to 2 decimals first (about 1 km). That is enough for nearest. You then never send your user's exact position, and the same area gives the same URL, so the day-long cache of the answer is used (see Limits and caching).
3. Ask for the answer. Call /now with the slug, then /days.
4. Show the credits. Wherever you show our data, show the credits that /v1/attribution lists. See Attribution.
5. Call from a server. A key in a web page or a mobile app is public to everyone who opens it. Keep it on your server and call us from there.

Set your key in the shell, then run the calls below. Each call shows the answer we recorded for it.

export SLEEVECAST_API_KEY=swk1_...your key...

Example: The city nearest to a point

curl -sS 'https://api.sleevecast.app/v1/cities/nearest?lat=-6.16&lon=39.19' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "zanzibar",
    "geonames_id": 148730,
    "name": "Zanzibar",
    "country": "TZ",
    "lat": -6.16394,
    "lon": 39.19793,
    "tz": "Africa/Dar_es_Salaam",
    "admin1": "25",
    "population": 403658,
    "cameras": false
  },
  "distance_km": 1.0
}

Example: A city with camera evidence

curl -sS 'https://api.sleevecast.app/v1/cities/madrid/now' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "madrid",
    "geonames_id": 3117735,
    "name": "Madrid",
    "country": "ES",
    "lat": 40.4165,
    "lon": -3.70256,
    "tz": "Europe/Madrid"
  },
  "at": "2026-10-26T14:00:00+01:00",
  "summary": "Mixed: short sleeves 36%, light jackets 26%, closed shoes",
  "fields": {
    "upper": {
      "sleeveless": 0.0,
      "short_sleeve": 0.36,
      "long_sleeve": 0.26,
      "knit_hoodie": 0.12,
      "light_jacket": 0.26,
      "coat": 0.0,
      "puffer": 0.0
    },
    "lower": {
      "shorts": 0.04,
      "skirt_short": 0.02,
      "skirt_long": 0.02,
      "trousers": 0.92
    },
    "footwear": {
      "open": 0.01,
      "closed_low": 0.98,
      "boots": 0.01
    },
    "scarf": {
      "yes": 0.0,
      "no": 1.0
    },
    "umbrella_open": {
      "yes": 0.0,
      "no": 1.0
    }
  },
  "index": 38,
  "basis": "model+cameras",
  "confidence": "high",
  "data_through": "2026-10-26T14:10:00+01:00",
  "weather": {
    "t": 17.0,
    "feels": 15.0,
    "precip": 0.0,
    "wind": 10.0,
    "cloud": 50,
    "code": 1,
    "issued_at": "2026-10-26T13:20:00+01:00",
    "units": {
      "t": "°C",
      "feels": "°C",
      "precip": "mm",
      "wind": "km/h",
      "cloud": "%"
    }
  },
  "sources": [
    "madrid-informo",
    "met-norway",
    "geonames"
  ]
}

Example: Three days in Madrid

curl -sS 'https://api.sleevecast.app/v1/cities/madrid/days?days=3' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "madrid",
    "geonames_id": 3117735,
    "name": "Madrid",
    "country": "ES",
    "lat": 40.4165,
    "lon": -3.70256,
    "tz": "Europe/Madrid"
  },
  "days": [
    {
      "date": "2026-10-26",
      "summary": "Mixed: short sleeves 36%, light jackets 26%, closed shoes",
      "fields": {
        "upper": {
          "sleeveless": 0.0,
          "short_sleeve": 0.36,
          "long_sleeve": 0.26,
          "knit_hoodie": 0.12,
          "light_jacket": 0.26,
          "coat": 0.0,
          "puffer": 0.0
        },
        "lower": {
          "shorts": 0.04,
          "skirt_short": 0.02,
          "skirt_long": 0.02,
          "trousers": 0.92
        },
        "footwear": {
          "open": 0.01,
          "closed_low": 0.98,
          "boots": 0.01
        },
        "scarf": {
          "yes": 0.0,
          "no": 1.0
        },
        "umbrella_open": {
          "yes": 0.0,
          "no": 1.0
        }
      },
      "index": 38,
      "basis": "model+cameras",
      "confidence": "high",
      "weather": {
        "t_min": 9.0,
        "t_max": 21.0,
        "feels": 15.0,
        "precip": 0.0,
        "wind": 10.0,
        "issued_at": "2026-10-26T13:20:00+01:00"
      }
    },
    {
      "date": "2026-10-27",
      "summary": "Mixed: light jackets 32%, sweaters and hoodies 24%, closed shoes",
      "fields": {
        "upper": {
          "sleeveless": 0.0,
          "short_sleeve": 0.08,
          "long_sleeve": 0.23,
          "knit_hoodie": 0.24,
          "light_jacket": 0.32,
          "coat": 0.09,
          "puffer": 0.04
        },
        "lower": {
          "shorts": 0.04,
          "skirt_short": 0.02,
          "skirt_long": 0.02,
          "trousers": 0.92
        },
        "footwear": {
          "open": 0.02,
          "closed_low": 0.93,
          "boots": 0.05
        },
        "scarf": {
          "yes": 0.01,
          "no": 0.99
        },
        "umbrella_open": {
          "yes": 0.0,
          "no": 1.0
        }
      },
      "index": 54,
      "basis": "model",
      "confidence": "medium",
      "weather": {
        "t_min": 17.0,
        "t_max": 17.0,
        "feels": 15.0,
        "precip": 0.0,
        "wind": 10.0,
        "issued_at": "2026-10-26T13:20:00+01:00"
      }
    },
    {
      "date": "2026-10-28",
      "summary": "Mixed: light jackets 28%, coats 27%, closed shoes; umbrellas 30%",
      "fields": {
        "upper": {
          "sleeveless": 0.0,
          "short_sleeve": 0.02,
          "long_sleeve": 0.09,
          "knit_hoodie": 0.16,
          "light_jacket": 0.28,
          "coat": 0.27,
          "puffer": 0.18
        },
        "lower": {
          "shorts": 0.02,
          "skirt_short": 0.01,
          "skirt_long": 0.02,
          "trousers": 0.95
        },
        "footwear": {
          "open": 0.01,
          "closed_low": 0.87,
          "boots": 0.12
        },
        "scarf": {
          "yes": 0.08,
          "no": 0.92
        },
        "umbrella_open": {
          "yes": 0.3,
          "no": 0.7
        }
      },
      "index": 71,
      "basis": "model",
      "confidence": "medium",
      "weather": {
        "t_min": 11.0,
        "t_max": 11.0,
        "feels": 8.0,
        "precip": 12.0,
        "wind": 22.0,
        "issued_at": "2026-10-26T13:20:00+01:00"
      }
    }
  ],
  "sources": [
    "madrid-informo",
    "met-norway",
    "geonames"
  ]
}

Keys

• Send the key in one way only: the header Authorization: Bearer swk1_.... A key in a URL, in another header or in a cookie counts as no key. Never put it in a URL.
• If a key ever went into a URL, we remove keys we recognise from URLs before logging, but treat a key that was ever sent in a URL as leaked: it stays in your logs and in your users' browsers. Ask us for a new key.
• One key per app or web page. A key that leaks can then be revoked without touching your other apps.
• A key expires. From its expiry date (UTC), calls fail with api_key_expired. Ask for a new key before that day.
• The four 401 codes tell you what is wrong with a key: api_key_missing (no key sent), api_key_invalid (malformed, or not one we issued), api_key_expired and api_key_revoked. A 401 carries no RateLimit-* headers: with no accepted key there is no budget.
• A rate limit is per key. See Limits and caching.

Cities

• A city has a slug and a GeoNames id. Use either in the path: /v1/cities/madrid/now or /v1/cities/3117735/now. The list shows both.
• Store the slug. A slug does not change, and we never give it to another city. A city that leaves the catalogue is still found by its slug or id, and comes last in the list.
• Search the list with q. q matches the start of a name, ignoring case, accents and apostrophes. São finds Sao Paulo, Koln and Koeln find Köln, and Xian finds Xi'an. It also matches the start of the name's plain-ASCII spelling and of the slug. It does not search other names: Moskva and München find nothing.
• Page through the list with limit and cursor. Send back the next_cursor of a page as the cursor of the next one. It is null on the last page. The most populous cities come first. Filter by country with a two-letter code.
• nearest finds the nearest city within 50 km of a point, or answers 404 with the code no_city_nearby. It answers with a city, not with one of its districts. A district is still in the list and answers like any city. Round the coordinates to 2 decimals before you send them: it protects your users, and it makes the day-long cache of the answer useful.
• cameras: true marks a city with our street cameras: what they see can correct its answers, field by field, once their labels pass our accuracy check. basis says whether an answer was corrected. Where it is false, the weather model answers alone.

Endpoints

All seven operations are GET. HEAD works too. Each is described below with its parameters, its answers and a recorded example.

• GET /v1/cities/{city}/now: what people wear in the city this hour. Add ?meta=true for diagnostics.
• GET /v1/cities/{city}/days: today and the days after it, up to 7 days in all. Use it for packing.
• GET /v1/cities: list or search the cities.
• GET /v1/cities/nearest: the city nearest to a point.
• GET /v1/cities/{city}/history: what our cameras saw, day by day. A city without cameras has no items. It shows the shares of a field only for a finished day, only for a field whose camera labels have passed our accuracy check, and only from 10 people. The default range is the last 30 days. The longest range is 366 days; a longer one is a 422.
• GET /v1/attribution: the credits to show.
• GET /v1/version: the versions behind the numbers.

GET /v1/cities/{city}/now

What people wear now

What people wear in the city now: the shares of each value of each field for the current local hour, from the weather model, which the city's licence-clean cameras can correct (basis says when).

Parameters

Responses

Example: No key sent

curl -sS 'https://api.sleevecast.app/v1/cities/madrid/now'

Answer: 401

{
  "error": {
    "code": "api_key_missing",
    "message": "Send the API key as 'Authorization: Bearer <key>'",
    "request_id": "5f0c3e9a2b7d4e18a6c1d0f2b3e4a5c6"
  }
}

Example: The key's budget is spent

curl -sS 'https://api.sleevecast.app/v1/cities/madrid/now' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 429

{
  "error": {
    "code": "rate_limited",
    "message": "Too many requests for this API key: 60 a minute; retry after 60 s",
    "request_id": "5f0c3e9a2b7d4e18a6c1d0f2b3e4a5c6"
  }
}

Example: No weather to answer with

curl -sS 'https://api.sleevecast.app/v1/cities/tokyo/now' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 503

{
  "error": {
    "code": "weather_unavailable",
    "message": "no stored forecast covers this hour",
    "request_id": "5f0c3e9a2b7d4e18a6c1d0f2b3e4a5c6"
  }
}

Example: A city with camera evidence

curl -sS 'https://api.sleevecast.app/v1/cities/madrid/now' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "madrid",
    "geonames_id": 3117735,
    "name": "Madrid",
    "country": "ES",
    "lat": 40.4165,
    "lon": -3.70256,
    "tz": "Europe/Madrid"
  },
  "at": "2026-10-26T14:00:00+01:00",
  "summary": "Mixed: short sleeves 36%, light jackets 26%, closed shoes",
  "fields": {
    "upper": {
      "sleeveless": 0.0,
      "short_sleeve": 0.36,
      "long_sleeve": 0.26,
      "knit_hoodie": 0.12,
      "light_jacket": 0.26,
      "coat": 0.0,
      "puffer": 0.0
    },
    "lower": {
      "shorts": 0.04,
      "skirt_short": 0.02,
      "skirt_long": 0.02,
      "trousers": 0.92
    },
    "footwear": {
      "open": 0.01,
      "closed_low": 0.98,
      "boots": 0.01
    },
    "scarf": {
      "yes": 0.0,
      "no": 1.0
    },
    "umbrella_open": {
      "yes": 0.0,
      "no": 1.0
    }
  },
  "index": 38,
  "basis": "model+cameras",
  "confidence": "high",
  "data_through": "2026-10-26T14:10:00+01:00",
  "weather": {
    "t": 17.0,
    "feels": 15.0,
    "precip": 0.0,
    "wind": 10.0,
    "cloud": 50,
    "code": 1,
    "issued_at": "2026-10-26T13:20:00+01:00",
    "units": {
      "t": "°C",
      "feels": "°C",
      "precip": "mm",
      "wind": "km/h",
      "cloud": "%"
    }
  },
  "sources": [
    "madrid-informo",
    "met-norway",
    "geonames"
  ]
}

Example: The same, with meta=true

curl -sS 'https://api.sleevecast.app/v1/cities/madrid/now?meta=true' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "madrid",
    "geonames_id": 3117735,
    "name": "Madrid",
    "country": "ES",
    "lat": 40.4165,
    "lon": -3.70256,
    "tz": "Europe/Madrid"
  },
  "at": "2026-10-26T14:00:00+01:00",
  "summary": "Mixed: short sleeves 36%, light jackets 26%, closed shoes",
  "fields": {
    "upper": {
      "sleeveless": 0.0,
      "short_sleeve": 0.36,
      "long_sleeve": 0.26,
      "knit_hoodie": 0.12,
      "light_jacket": 0.26,
      "coat": 0.0,
      "puffer": 0.0
    },
    "lower": {
      "shorts": 0.04,
      "skirt_short": 0.02,
      "skirt_long": 0.02,
      "trousers": 0.92
    },
    "footwear": {
      "open": 0.01,
      "closed_low": 0.98,
      "boots": 0.01
    },
    "scarf": {
      "yes": 0.0,
      "no": 1.0
    },
    "umbrella_open": {
      "yes": 0.0,
      "no": 1.0
    }
  },
  "index": 38,
  "basis": "model+cameras",
  "confidence": "high",
  "data_through": "2026-10-26T14:10:00+01:00",
  "weather": {
    "t": 17.0,
    "feels": 15.0,
    "precip": 0.0,
    "wind": 10.0,
    "cloud": 50,
    "code": 1,
    "issued_at": "2026-10-26T13:20:00+01:00",
    "units": {
      "t": "°C",
      "feels": "°C",
      "precip": "mm",
      "wind": "km/h",
      "cloud": "%"
    }
  },
  "sources": [
    "madrid-informo",
    "met-norway",
    "geonames"
  ],
  "meta": {
    "model_version": "m1.b804897b",
    "rules_version": "a1.3c5f046b",
    "catalog_version": "gn15k.8f3e7ea7",
    "labeler": "openrouter:google/gemini-3.1-flash-lite",
    "taxonomy_version": "v1",
    "unit": {
      "key": "iw:catholic_europe",
      "calibrated": true,
      "shift_c": -1.5
    },
    "prior_strength": 20.0,
    "window": {
      "from": "2026-10-26T12:00:00+01:00",
      "hours": 3,
      "n_cameras": 1,
      "n": {
        "upper": 23,
        "lower": 9,
        "footwear": 23,
        "scarf": 23,
        "umbrella_open": 23
      },
      "camera_weight": {
        "upper": 0.53,
        "lower": 0.0,
        "footwear": 0.53,
        "scarf": 0.53,
        "umbrella_open": 0.53
      }
    },
    "weather_point": {
      "slug": "madrid",
      "distance_km": 0.0
    }
  }
}

Example: A city the model answers alone

curl -sS 'https://api.sleevecast.app/v1/cities/sydney/now' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "sydney",
    "geonames_id": 2147714,
    "name": "Sydney",
    "country": "AU",
    "lat": -33.86785,
    "lon": 151.20732,
    "tz": "Australia/Sydney"
  },
  "at": "2026-10-26T15:00:00+11:00",
  "summary": "Mixed: long sleeves 33%, light jackets 22%, closed shoes",
  "fields": {
    "upper": {
      "sleeveless": 0.01,
      "short_sleeve": 0.18,
      "long_sleeve": 0.33,
      "knit_hoodie": 0.21,
      "light_jacket": 0.22,
      "coat": 0.04,
      "puffer": 0.01
    },
    "lower": {
      "shorts": 0.07,
      "skirt_short": 0.02,
      "skirt_long": 0.03,
      "trousers": 0.88
    },
    "footwear": {
      "open": 0.02,
      "closed_low": 0.94,
      "boots": 0.04
    },
    "scarf": {
      "yes": 0.01,
      "no": 0.99
    },
    "umbrella_open": {
      "yes": 0.0,
      "no": 1.0
    }
  },
  "index": 44,
  "basis": "model",
  "confidence": "low",
  "data_through": null,
  "weather": {
    "t": 19.4,
    "feels": 18.3,
    "precip": 0.0,
    "wind": 17.1,
    "cloud": 11,
    "code": 0,
    "issued_at": "2026-10-26T14:05:00+11:00",
    "units": {
      "t": "°C",
      "feels": "°C",
      "precip": "mm",
      "wind": "km/h",
      "cloud": "%"
    }
  },
  "sources": [
    "met-norway",
    "geonames"
  ]
}

GET /v1/cities/{city}/days

What people wear today and on the next days

What people wear in the city today and on the days after it, for packing: one item per local date, today first, over the day's outdoor hours. Today's item can carry the cameras' correction where the city has licence-clean cameras; later days are the weather model's.

Parameters

Responses

Example: Three days in Madrid

curl -sS 'https://api.sleevecast.app/v1/cities/madrid/days?days=3' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "madrid",
    "geonames_id": 3117735,
    "name": "Madrid",
    "country": "ES",
    "lat": 40.4165,
    "lon": -3.70256,
    "tz": "Europe/Madrid"
  },
  "days": [
    {
      "date": "2026-10-26",
      "summary": "Mixed: short sleeves 36%, light jackets 26%, closed shoes",
      "fields": {
        "upper": {
          "sleeveless": 0.0,
          "short_sleeve": 0.36,
          "long_sleeve": 0.26,
          "knit_hoodie": 0.12,
          "light_jacket": 0.26,
          "coat": 0.0,
          "puffer": 0.0
        },
        "lower": {
          "shorts": 0.04,
          "skirt_short": 0.02,
          "skirt_long": 0.02,
          "trousers": 0.92
        },
        "footwear": {
          "open": 0.01,
          "closed_low": 0.98,
          "boots": 0.01
        },
        "scarf": {
          "yes": 0.0,
          "no": 1.0
        },
        "umbrella_open": {
          "yes": 0.0,
          "no": 1.0
        }
      },
      "index": 38,
      "basis": "model+cameras",
      "confidence": "high",
      "weather": {
        "t_min": 9.0,
        "t_max": 21.0,
        "feels": 15.0,
        "precip": 0.0,
        "wind": 10.0,
        "issued_at": "2026-10-26T13:20:00+01:00"
      }
    },
    {
      "date": "2026-10-27",
      "summary": "Mixed: light jackets 32%, sweaters and hoodies 24%, closed shoes",
      "fields": {
        "upper": {
          "sleeveless": 0.0,
          "short_sleeve": 0.08,
          "long_sleeve": 0.23,
          "knit_hoodie": 0.24,
          "light_jacket": 0.32,
          "coat": 0.09,
          "puffer": 0.04
        },
        "lower": {
          "shorts": 0.04,
          "skirt_short": 0.02,
          "skirt_long": 0.02,
          "trousers": 0.92
        },
        "footwear": {
          "open": 0.02,
          "closed_low": 0.93,
          "boots": 0.05
        },
        "scarf": {
          "yes": 0.01,
          "no": 0.99
        },
        "umbrella_open": {
          "yes": 0.0,
          "no": 1.0
        }
      },
      "index": 54,
      "basis": "model",
      "confidence": "medium",
      "weather": {
        "t_min": 17.0,
        "t_max": 17.0,
        "feels": 15.0,
        "precip": 0.0,
        "wind": 10.0,
        "issued_at": "2026-10-26T13:20:00+01:00"
      }
    },
    {
      "date": "2026-10-28",
      "summary": "Mixed: light jackets 28%, coats 27%, closed shoes; umbrellas 30%",
      "fields": {
        "upper": {
          "sleeveless": 0.0,
          "short_sleeve": 0.02,
          "long_sleeve": 0.09,
          "knit_hoodie": 0.16,
          "light_jacket": 0.28,
          "coat": 0.27,
          "puffer": 0.18
        },
        "lower": {
          "shorts": 0.02,
          "skirt_short": 0.01,
          "skirt_long": 0.02,
          "trousers": 0.95
        },
        "footwear": {
          "open": 0.01,
          "closed_low": 0.87,
          "boots": 0.12
        },
        "scarf": {
          "yes": 0.08,
          "no": 0.92
        },
        "umbrella_open": {
          "yes": 0.3,
          "no": 0.7
        }
      },
      "index": 71,
      "basis": "model",
      "confidence": "medium",
      "weather": {
        "t_min": 11.0,
        "t_max": 11.0,
        "feels": 8.0,
        "precip": 12.0,
        "wind": 22.0,
        "issued_at": "2026-10-26T13:20:00+01:00"
      }
    }
  ],
  "sources": [
    "madrid-informo",
    "met-norway",
    "geonames"
  ]
}

Example: One day in Sydney, with meta=true

curl -sS 'https://api.sleevecast.app/v1/cities/sydney/days?days=1&meta=true' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "sydney",
    "geonames_id": 2147714,
    "name": "Sydney",
    "country": "AU",
    "lat": -33.86785,
    "lon": 151.20732,
    "tz": "Australia/Sydney"
  },
  "days": [
    {
      "date": "2026-10-26",
      "summary": "Mixed: long sleeves 33%, light jackets 22%, closed shoes",
      "fields": {
        "upper": {
          "sleeveless": 0.01,
          "short_sleeve": 0.18,
          "long_sleeve": 0.33,
          "knit_hoodie": 0.21,
          "light_jacket": 0.22,
          "coat": 0.04,
          "puffer": 0.01
        },
        "lower": {
          "shorts": 0.07,
          "skirt_short": 0.02,
          "skirt_long": 0.03,
          "trousers": 0.88
        },
        "footwear": {
          "open": 0.02,
          "closed_low": 0.94,
          "boots": 0.04
        },
        "scarf": {
          "yes": 0.01,
          "no": 0.99
        },
        "umbrella_open": {
          "yes": 0.0,
          "no": 1.0
        }
      },
      "index": 44,
      "basis": "model",
      "confidence": "low",
      "weather": {
        "t_min": 19.4,
        "t_max": 19.4,
        "feels": 18.3,
        "precip": 0.0,
        "wind": 17.1,
        "issued_at": "2026-10-26T14:05:00+11:00"
      }
    }
  ],
  "sources": [
    "met-norway",
    "geonames"
  ],
  "meta": {
    "model_version": "m1.b804897b",
    "rules_version": "a1.3c5f046b",
    "catalog_version": "gn15k.8f3e7ea7",
    "labeler": "openrouter:google/gemini-3.1-flash-lite",
    "taxonomy_version": "v1",
    "unit": {
      "key": null,
      "calibrated": false,
      "shift_c": -1.5
    },
    "prior_strength": 20.0,
    "window": null,
    "weather_point": {
      "slug": "sydney",
      "distance_km": 0.0
    }
  }
}

Example: A parameter out of range

curl -sS 'https://api.sleevecast.app/v1/cities/madrid/days?days=8' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 422

{
  "error": {
    "code": "validation_error",
    "message": "query.days: Input should be less than or equal to 7",
    "request_id": "5f0c3e9a2b7d4e18a6c1d0f2b3e4a5c6"
  }
}

GET /v1/cities

List or search cities

The cities of the catalogue, a page at a time, the most populous first; cities that have left the catalogue come last. Without q it lists them all, with it it searches by name. The next_cursor of a page goes back as the cursor of the next one, and is null on the last.

Parameters

Responses

Example: The first page of a country

curl -sS 'https://api.sleevecast.app/v1/cities?country=ES&limit=1' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "items": [
    {
      "slug": "madrid",
      "geonames_id": 3117735,
      "name": "Madrid",
      "country": "ES",
      "lat": 40.4165,
      "lon": -3.70256,
      "tz": "Europe/Madrid",
      "admin1": "29",
      "population": 3255944,
      "cameras": true
    }
  ],
  "next_cursor": "eyJyIjowLCJwIjozMjU1OTQ0LCJzIjoibWFkcmlkIn0"
}

Example: Search by name

curl -sS 'https://api.sleevecast.app/v1/cities?q=sao' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "items": [
    {
      "slug": "sao-paulo",
      "geonames_id": 3448439,
      "name": "São Paulo",
      "country": "BR",
      "lat": -23.5475,
      "lon": -46.63611,
      "tz": "America/Sao_Paulo",
      "admin1": "27",
      "population": 10021295,
      "cameras": false
    }
  ],
  "next_cursor": null
}

GET /v1/cities/nearest

The city nearest to a point

The city nearest to a point, with its distance from it in km, for a client that knows where it is and needs a city to ask now and days about. A point far from every city (the open sea, a desert) is a 404 no_city_nearby, whose message says how far counts as near. The answer changes only when the catalogue is reloaded, so the client may keep it for a day.

Parameters

Responses

Example: A point with no city nearby

curl -sS 'https://api.sleevecast.app/v1/cities/nearest?lat=0&lon=-140' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 404

{
  "error": {
    "code": "no_city_nearby",
    "message": "no city within 50 km",
    "request_id": "5f0c3e9a2b7d4e18a6c1d0f2b3e4a5c6"
  }
}

Example: The city nearest to a point

curl -sS 'https://api.sleevecast.app/v1/cities/nearest?lat=-6.16&lon=39.19' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "zanzibar",
    "geonames_id": 148730,
    "name": "Zanzibar",
    "country": "TZ",
    "lat": -6.16394,
    "lon": 39.19793,
    "tz": "Africa/Dar_es_Salaam",
    "admin1": "25",
    "population": 403658,
    "cameras": false
  },
  "distance_km": 1.0
}

GET /v1/cities/{city}/history

What our cameras saw, day by day

What the cameras of a city saw, one item per local day that has evidence: the people judged on each field, the day's weather and, for the fields whose camera labels have been verified, once the day is over and only when enough people were judged (the privacy floor), the shares of the field and the warmth index. A day in progress shows its counts and its weather and no shares. A city without cameras has no history: 200 with no items.

Parameters

Responses

Example: Three days of history

curl -sS 'https://api.sleevecast.app/v1/cities/madrid/history?from=2026-10-24&to=2026-10-26' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "city": {
    "slug": "madrid",
    "geonames_id": 3117735,
    "name": "Madrid",
    "country": "ES",
    "lat": 40.4165,
    "lon": -3.70256,
    "tz": "Europe/Madrid"
  },
  "rules_version": "a1.3c5f046b",
  "sources": [
    "madrid-informo"
  ],
  "items": [
    {
      "date": "2026-10-24",
      "complete": true,
      "data_through": "2026-10-24T15:10:00+02:00",
      "fields": {
        "upper": {
          "sleeveless": 0.0,
          "short_sleeve": 0.0,
          "long_sleeve": 0.0,
          "knit_hoodie": 0.0,
          "light_jacket": 0.53,
          "coat": 0.47,
          "puffer": 0.0
        },
        "lower": {
          "shorts": 0.0,
          "skirt_short": 0.0,
          "skirt_long": 0.0,
          "trousers": 1.0
        },
        "footwear": {
          "open": 0.0,
          "closed_low": 1.0,
          "boots": 0.0
        },
        "scarf": {
          "yes": 0.0,
          "no": 1.0
        },
        "umbrella_open": {
          "yes": 0.0,
          "no": 1.0
        }
      },
      "n": {
        "upper": 17,
        "lower": 17,
        "footwear": 17,
        "scarf": 17,
        "umbrella_open": 17
      },
      "index": 75,
      "weather": {
        "t_min": 9.0,
        "t_max": 17.0,
        "feels": 15.2
      }
    },
    {
      "date": "2026-10-25",
      "complete": true,
      "data_through": "2026-10-25T14:10:00+01:00",
      "fields": {
        "upper": {
          "sleeveless": 0.0,
          "short_sleeve": 0.0,
          "long_sleeve": 0.0,
          "knit_hoodie": 0.24,
          "light_jacket": 0.0,
          "coat": 0.53,
          "puffer": 0.23
        },
        "lower": {
          "shorts": 0.0,
          "skirt_short": 0.0,
          "skirt_long": 0.24,
          "trousers": 0.76
        },
        "footwear": {
          "open": 0.0,
          "closed_low": 0.82,
          "boots": 0.18
        },
        "umbrella_open": {
          "yes": 0.18,
          "no": 0.82
        }
      },
      "n": {
        "upper": 17,
        "lower": 17,
        "footwear": 17,
        "scarf": 9,
        "umbrella_open": 17
      },
      "index": 79,
      "weather": {
        "t_min": 6.0,
        "t_max": 14.0,
        "feels": 11.7
      }
    },
    {
      "date": "2026-10-26",
      "complete": false,
      "data_through": "2026-10-26T14:10:00+01:00",
      "fields": {},
      "n": {
        "upper": 17,
        "lower": 17,
        "footwear": 17,
        "scarf": 9,
        "umbrella_open": 17
      },
      "index": null,
      "weather": {
        "t_min": 9.0,
        "t_max": 17.0,
        "feels": null
      }
    }
  ]
}

GET /v1/attribution

The credits to show

The credits of the service, for a client to show once, on a credits page or behind a link, and not in every answer: the camera sources whose cameras or stored days the answers rest on (ordered by operator, then id; cities are the cities they serve) and the datasets the service is built on (the city names, the weather, the regions).

Responses

Example: The credits to show

curl -sS 'https://api.sleevecast.app/v1/attribution' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "sources": [
    {
      "id": "madrid-informo",
      "operator": "Ayuntamiento de Madrid",
      "licence": "CC BY 4.0",
      "attribution": "Origen de los datos: Ayuntamiento de Madrid",
      "terms_url": "https://datos.madrid.es/dataset/202088-0-trafico-camaras",
      "cities": [
        "madrid"
      ]
    },
    {
      "id": "austin-tpw",
      "operator": "City of Austin Transportation and Public Works",
      "licence": "Public Domain",
      "attribution": null,
      "terms_url": "https://data.austintexas.gov/d/b4k4-adkb",
      "cities": [
        "austin"
      ]
    }
  ],
  "datasets": [
    {
      "id": "geonames",
      "name": "GeoNames",
      "use": "city names, coordinates, population",
      "licence": "CC BY 4.0",
      "attribution": "GeoNames (geonames.org)",
      "url": "https://www.geonames.org/"
    },
    {
      "id": "open-meteo",
      "name": "Open-Meteo",
      "use": "weather forecasts and observations",
      "licence": "CC BY 4.0",
      "attribution": "Weather data by Open-Meteo.com",
      "url": "https://open-meteo.com/"
    },
    {
      "id": "met-norway",
      "name": "MET Norway Locationforecast 2.0",
      "use": "weather forecasts (hourly values interpolated and feels-like temperature computed by sleevecast)",
      "licence": "CC BY 4.0",
      "attribution": "Weather data from MET Norway",
      "url": "https://api.met.no/weatherapi/locationforecast/2.0/"
    },
    {
      "id": "natural-earth",
      "name": "Natural Earth",
      "use": "regions: cultural units and climate classes",
      "licence": "public domain",
      "attribution": "Made with Natural Earth",
      "url": "https://www.naturalearthdata.com/"
    },
    {
      "id": "un-m49",
      "name": "UN M49",
      "use": "regions: cultural units and climate classes",
      "licence": "UN site terms",
      "attribution": "United Nations Statistics Division, Standard country or area codes for statistical use (M49)",
      "url": "https://unstats.un.org/unsd/methodology/m49/overview/"
    }
  ]
}

GET /v1/version

The versions behind the numbers

The versions behind the numbers of the other routes: the contract, the build, the taxonomy and the labeler, the model, the rules and the city catalog. It is always revalidated (Cache-Control: no-cache), so after a release it names the new build at once.

Responses

Example: The versions behind the numbers

curl -sS 'https://api.sleevecast.app/v1/version' -H "Authorization: Bearer $SLEEVECAST_API_KEY"

Answer: 200

{
  "service": "sleevecast",
  "api": "v1",
  "contract_version": "1.0",
  "version": "0.1.0",
  "build": "ec0854f08f933c0395b40594561948a3725e7551",
  "taxonomy_version": "v1",
  "labeler": "openrouter:google/gemini-3.1-flash-lite",
  "model_version": "m1.b804897b",
  "rules_version": "a1.3c5f046b",
  "catalog_version": "gn15k.8f3e7ea7"
}

Reading an answer

Shares

• Each field is a set of shares. fields.upper has one number for each upper-layer value. Each number is the expected share of the people outside in this city who wear that value. Shares have two decimals and add up to 1 as decimals; a floating-point sum can be off by about 1e-15.
• To rank values, sort the numbers yourself. Do not rely on the order of the keys.
• summary is a short English phrase over the shares. Show it if you like. Do not parse it: its wording may change.
• index is the warmth index, from 0 (everyone in sleeveless tops) to 100 (everyone in a puffer). It is side information.
• The values of each field are listed under Field values.

Where an answer comes from

• basis is model when the weather model made the answer alone, and model+cameras when what our cameras saw corrected it. Cameras can enter field by field, once their labels have passed our accuracy check.
• confidence says how far to trust the answer: high, medium or low. The schemas say what each level means. Show high and medium plainly. For low, show the answer with a softer wording, such as "a rough guess".
• data_through is the time of the newest camera observation behind the answer, to the minute. It is in /now and in every /history item; the items of /days have none. It is null when basis is model.
• The privacy floor. Camera evidence enters a field only when at least 10 people were judged on it in the camera window. The n counts in meta and in /history are whole people. /history shows a field's shares only from 10 people.
• sources lists the ids of what the answer rests on. Each id is listed in /v1/attribution, with the credit to show.
• meta is there only when you ask for it with ?meta=true. It holds the versions, the region's calibration and the camera window. It is for diagnostics. Do not show it to people.

Weather, times and zones

• weather in /now holds the temperature (t), the apparent temperature (feels), precip, wind, cloud and the WMO weather code, with a units object. Temperatures are in °C, precipitation in mm, wind in km/h. A value the forecast lacks is null. In /days, t_min and t_max cover the whole date, and feels, precip and wind cover its daytime hours.
• issued_at is when we last refreshed the stored forecast for this city.
• Times carry an offset, for example 2026-10-26T14:00:00+01:00. at is the start of the local hour the answer describes. For anything local, use the zone in city.tz and not the offset.
• Dates in /days and /history are local dates of the city.
• Numbers change. The shares change as the model, its calibrations and our cameras change. That is not a change of the contract, and it has no changelog entry.

Field values

These are all the values of each field, with their meaning. A value never changes its meaning. A new value would come with a new taxonomy and a new version.

upper

lower

footwear

scarf

umbrella_open

Errors

An error that the API itself answers has the body ErrorOut: one property, error, with a code, a message and a request_id (see the example "No key sent" under GET /v1/cities/{city}/now). The code is stable and meant for your program. The message is for people, and its wording may change. The same request_id is in the X-Request-ID header. Errors are never cached.

• When to retry. After a 429 or a 503, wait as long as Retry-After says, then retry. After a 500, retry once, with backoff. Do not retry other errors: they fail again until you change the request.

• weather_unavailable means the forecast we hold does not cover the hours you asked for. It is rare. Retry after the time in Retry-After; asking for fewer days may work sooner.

• Not every error answer is ours. Some answers come from the proxy in front of the API, and they are not in the table above:

  • a 404 for a path outside /v1, /guide and /healthz;
  • a 429 when requests arrive too fast;
  • a 502, 503 or 504 when no API server is ready or one does not answer in time.

  These answers carry no X-Request-ID and no ErrorOut body, and a 502, 503 or 504 carries no Retry-After. Do not parse their body, and treat a 502, 503 or 504 without our JSON as "try again later" with your own backoff.

• A page of another origin cannot call /v1 until its origin is allowed. While none is (the default), its preflight gets our 405 method_not_allowed. Once origins are allowed, a preflight from another origin, or one asking for another method or header, gets a 400 with a short plain-text body (from another origin, without Access-Control-Allow-Origin). Either way the browser reports a CORS error. Origins are allowed one by one, on request: see Support.

Limits and caching

The numbers below are the settings of the server that sent you this page.

• Per key. A key may make the number of requests in the table in any 60 seconds. Ask for more if you need it.
• RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset come on every answer to an accepted key (the number of seconds until the oldest counted request leaves the window). The 429 for a spent budget carries Retry-After. A refused request does not count against the budget. A 304 and a 404 do. When we run more than one server the counts are approximate.
• Failed key checks are limited per client address and overall; a valid key is never refused for them.
• At the proxy. A proxy in front of the API also limits how fast requests may arrive, to keep floods out. Its 429 is not an ErrorOut and has no RateLimit-* headers. It may carry Retry-After: honour it, or else wait a few seconds.
• Cache what the headers allow. An answer carries Cache-Control: max-age= with the seconds in the table, and a weak ETag. Within that time, ask again only if you must. Afterwards send If-None-Match with the ETag: you get a 304 with no body when the answer has not changed. A 304 still counts against your budget.
• nearest is kept for a day. It carries Cache-Control: private, max-age=86400. private means that your client alone keeps it, because its URL holds a position. A cache stores an answer by URL, so round the coordinates to 2 decimals: with 12.3456, the next call from the same phone a few metres away is a new URL, and the cache does not help.
• Poll gently. A city's answer changes at most every 30 minutes (cameras) and every hour (weather): poll /now at most every 5 minutes per city.
• /v1/version revalidates. It carries Cache-Control: no-cache, so you always see a new release at once; the ETag makes it a 304 when nothing changed.
• Send Accept-Encoding: gzip. Answers of 1,000 bytes or more are compressed.
• Browser origins are allowed one by one, on request. By default none is allowed, and a page of another origin cannot read /v1. Write to the address under Support with the origin.

Attribution

• The duty. Wherever you show data from this API, show the credits that /v1/attribution lists, or link to a page of yours that shows them. The licences of the data (CC BY 4.0 for MET Norway, GeoNames and Madrid's cameras) ask for credit, and CC BY 4.0 accepts a link to a page that shows it. This is also condition 3 of the conditions of use.
• The credits are two lists. sources lists the camera sources the answers can rest on: who operates the cameras, the licence, the credit line to show and the page of the terms. datasets lists the data the service is built on, such as GeoNames and the weather from MET Norway. Each entry names its licence and has a credit line to show.
• Show what an entry has. A camera source may list no city, and may have null for its licence, credit line or terms page. Show what it has.
• Keep the list current. New sources appear. Refresh it at least once a week. The sources of an answer name what that answer rests on, and each id is in the list.
• You need not show them in every answer. A credits page linked from where you show the data (for example behind an "i" button) is enough.

Conditions of use

1. Your key. A key is issued to one organisation for the use agreed when it was issued. Keep it out of public code, and use one key per app or web page. We may suspend or revoke a key that is abused or leaked, and we tell you why.
2. Fair use. Stay within your key's budget and cache what the headers allow. Do not spread one app's traffic over several keys to get around a limit; ask if you need more.
3. Credits. Show the credits that /v1/attribution lists wherever you show data from this API, or on a page linked from there. They are the conditions of the data's licences (CC BY 4.0 for MET Norway, GeoNames and Madrid's cameras), not ours.
4. The data. Weather values, place names and coordinates come under the licences in /v1/attribution; these conditions add nothing to them. The clothing estimates are ours: you may show them in your app and store them for that. Ask us before you redistribute them as a dataset.
5. Estimates. Answers are statistical estimates, and confidence says how far to trust each one. Do not rely on them where a wrong answer could cause harm.
6. No warranty. The service is offered as it is, without a guaranteed uptime, during a pilot. We change it only as the versioning rules allow.
7. Personal data. Do not send us personal data: name cities by their slugs, and round coordinates to two decimals.
8. Changes. We announce changes to these conditions in the changelog at least 30 days before they apply.

Operator: sleevecast.app.

Contact: api@sleevecast.app

Privacy

Requests this API answers are logged for abuse handling, with the IP address they came from, the method, the path and query string and the response status, and for /v1 also the name of the key (never the key itself) and the time taken; our log store keeps these lines for 7 days, and the servers that run the API keep a rolling copy, which new lines overwrite and which is deleted when the API is redeployed. A /v1 line also holds the host name the request was sent to and a request id, and quotes coordinates rounded to two decimals; no line holds another request header or a cookie. The guide pages set no cookie and load nothing from other sites. Requests to /v1 are also counted per key name, without addresses. Questions about these logs: write to the contact under Support.

Versioning

• The contract version is MAJOR.MINOR. The major is the path, /v1. The minor rises with each release that adds something. It is in the document (info.version), in GET /v1/version and in the Changelog.
• What we promise. Within /v1, additions only: a new property, a new code, a new operation, a new optional parameter. We do not remove or rename anything, and a value keeps its meaning.
• What you must do.
  • Ignore properties you do not know.
  • Accept error codes, source and dataset ids, and WMO weather codes you do not know.
  • Do not rely on the order of keys.
  • Do not parse summary or message.
  • Follow Retry-After.
• What is not a change of the contract. The numbers in the answers, the catalogue of cities, the camera sources, and the wording of summary, message, descriptions and examples. They change without a changelog entry.
• A change that would break a client goes to /v2. The one exception is a correction: the document was wrong about what the server always did. A correction ships with a changelog entry that names each change it makes.
• A deprecation within /v1 marks the property or parameter deprecated in the document and adds a changelog entry. It stays until /v2.
• /v2. If we ever ship it, /v1 stays served for at least 90 days after that. From the announcement on, every /v1 answer carries the headers Deprecation and Sunset. We also write to the contact of every key.

Changelog

This is the changelog of the public API under /v1. Within /v1 the contract only grows: a new property, a new code or a new operation is an addition. A change that would break a client goes to /v2. The one exception is a correction: the document was wrong about what the server always did, and its entry names each change it makes.

Each release has an entry below, newest first, with its date, the digest of the contract it describes, and what changed. Numbers, the catalogue, sources and the wording of summary, message, descriptions and examples change without an entry: they are not part of the contract. Changes to the conditions of use are announced here at least 30 days before they apply, as a dated paragraph above the entries.

1.0 — 2026-11-04

Contract 7676523d6b50734e, first.

• The seven operations: GET /v1/cities, /v1/cities/nearest, /v1/cities/{city}/now, /v1/cities/{city}/days, /v1/cities/{city}/history, /v1/attribution and /v1/version.
• The first frozen contract; before it, /v1 changed in place while no client held a key.

Semantics: as served since 2026-11-03, except that times carry their seconds (RFC 3339) and end in Z at a zero offset, /v1/version adds contract_version, and some properties are listed in a new order (a field's shares in taxonomy order, admin1 after tz).

Support

Contact: api@sleevecast.app

• Questions, a new key, a new origin, more budget: write to the address above.
• When you report a problem, include the request_id of the error body or the X-Request-ID header, the time of the call (with its zone), and the name of your key. Never send the key itself.
• The same address is for the operators of the services we use. MET Norway's requests name this page.

Reference: schemas

These are the bodies of the API, with every property and its meaning. They come from the same document as https://api.sleevecast.app/guide/openapi.json.

AttributionOut

The credits of the service, for a client to show once, on a credits page or behind a link, and not in every answer.

Basis

How an answer was made: from the weather model alone (model), or from the model corrected by what the cameras saw (model+cameras). A new value would be a breaking change.

Type: string: model, model+cameras

CameraSource

A source of camera images the answers rest on, to credit: who operates the cameras, the licence they publish under, the credit line to show and the page of their terms.

CameraWindow

The window the cameras' evidence came from: its first local hour, its length in hours, the number of cameras seen, and per field the people judged and the cameras' weight. No camera is named, and no share of the cameras alone is shown.

CitiesOut

A page of the catalogue: the most populous city first, and the cities that have left the catalogue last.

City

A city of the catalogue: what CityRef names and locates, with its first-level division, its population, and whether our cameras can correct its answers.

CityRef

The city an answer is for: what a client needs to name and locate it.

Confidence

How far to trust an answer, from low to high. A new value would be a breaking change.

Type: string: low, medium, high

Dataset

A dataset the service is built on, to credit.

Day

The answer for one local date: the shares of each field over the day's outdoor hours, the warmth index of upper, how the answer was made and the day's weather. Only the first item, today, can have basis model+cameras.

DayWeather

The weather of a local date, to one decimal. t_min and t_max are the lowest and highest hourly temperature of the whole date, feels and wind the means and precip the total over its daytime hours. Only the hours the forecast covers count, and a value that none of the hours it is taken over has is null.

DaysOut

The answers for today and the days after it, today first. meta is there only when ?meta=true was asked, once for the whole response.

ErrorBody

What went wrong: a stable code, a message for people and the id of the request.

ErrorOut

The body of an error that the API itself answers under /v1, whatever its status. The proxy in front of the API has answers of its own (a 404 for another path, a 429, a 502, 503 or 504), and those have no such body.

FieldCounts

The people the cameras judged on each field: whole people, counts and never shares.

FieldWeights

The weight of the cameras' evidence in the answer for each field, from 0 (the model alone) to 1.

Fields

The shares of each clothing field the service answers for: every value of every field is a named property, always there.

FootwearShares

The shares of footwear, what is worn on the feet: one property per value, from 0 to 1 with two decimals, in the taxonomy's order; they sum to 1. Closed: a new value would be a new taxonomy.

HistoryDay

What the cameras observed on one local day. n, data_through and weather are always shown. fields and index are shown only when all of these hold: the day is over; the field's camera labels have passed our accuracy check; and enough people were judged on the field that day. A day in progress has fields empty and index null. A day can still change for a while after it is over, as late labels arrive, so a client must not treat it as final.

HistoryFields

The shares the cameras observed on one day, for the fields that are shown. A field is there only for a finished day, only if its camera labels have passed our accuracy check, and only when enough people were judged on it that day (the privacy floor); otherwise it is absent, never null. Today's item has none.

HistoryOut

The history of a city: one item per local day with camera evidence in the range asked, oldest first. A city without cameras has none.

HistoryWeather

The weather of a day in the history, to one decimal. A value the day has no data for is null.

LowerShares

The shares of lower, what is worn on the legs: one property per value, from 0 to 1 with two decimals, in the taxonomy's order; they sum to 1. Closed: a new value would be a new taxonomy.

Meta

What ?meta=true adds to an answer: the versions it was made with, the city's unit and shift, the cameras' window and the weather point.

ModelUnit

The cultural unit of the city, whether a calibration names it, and the shift in °C its feels-like temperature is read at.

NearestOut

The city nearest to a point and its distance from it.

NowOut

The answer for a city's current local hour. meta is there only when ?meta=true was asked.

NowWeather

The weather of the hour and the forecast it comes from. t, feels, precip and wind are to one decimal. A value the forecast lacks for the hour is null.

UpperShares

The shares of upper, the outermost layer on the upper body: one property per value, from 0 to 1 with two decimals, in the taxonomy's order; they sum to 1. Closed: a new value would be a new taxonomy.

VersionOut

The contract, the build, the labeler and the rules behind the numbers of the other routes.

WeatherPoint

The city whose stored forecast the answer read.

WeatherUnits

The units of NowWeather's values.

YesNoShares

The shares of a yes-or-no field, scarf or umbrella_open: yes and no, from 0 to 1 with two decimals; they sum to 1. Closed: a new value would be a new taxonomy.