Yahoo Weather Scraper

Collect current conditions and up to eight days of Yahoo Weather forecasts for one or many public locations. Version 3 uses Yahoo's server-rendered structured React payload, so it needs one lightweight HTTP request per location and does not wait for a browser to render advertisements or client scripts.

Input

locations accepts user-friendly Town, REGION, COUNTRY values such as Nashville, TN, US and London, England, GB. Existing country/region/town paths and complete https://weather.yahoo.com/... URLs remain supported. Town and region names are converted to Yahoo's lowercase URL slugs; country must be the two-letter code used in Yahoo's route.

Use detailLevel to control output size:

• current: current conditions and weather detail metrics only.
• daily: current conditions plus daily temperature, precipitation, and wind summaries.
• full: daily output plus aligned hourly temperature, condition, precipitation, and wind periods.

forecastDays selects 1–8 daily entries and maxHourlyItems limits hourly entries per location. Current-temperature and condition filters, four sort modes, exact maxItems, bounded concurrency/retries/timeouts, direct access, Apify Residential proxies, and custom HTTP proxy URLs are supported. Direct access is the default. GOOGLE_SERP is rejected because it is not compatible with Yahoo HTTPS pages.

Output

Each dataset row represents one Yahoo Weather location. Stable location identity, geographic coordinates, typed current temperature, RealFeel, today's high/low, condition, sunrise/sunset, humidity, air quality, UV, allergens, pressure, visibility, moon phase, daily summaries, and hourly periods are included when published by Yahoo. Optional source fields are omitted when unavailable; the Actor never writes null, empty-string, challenge-page, or fabricated placeholder values.

Daily labels are stored exactly as Yahoo publishes them (Today, Tomorrow, Tue 1, and similar) with a numeric dayOffset. The page does not publish an unambiguous time zone in this payload, so the Actor does not fabricate ISO dates from those labels.

Reliability and access

The Actor validates HTTP status, HTML content type, final host, response size, structured payload shape, location identity, and required forecast fields before normalization. It uses sessions and bounded temporary retries. A CAPTCHA, verification page, unexpected redirect, or changed payload fails closed and no partial dataset is written. Proxies are for ordinary authorized access resilience, not for bypassing authentication, CAPTCHAs, subscriptions, or other access controls.

Example

{
  "locations": ["Nashville, TN, US", "Sacramento, CA, US", "New York, NY, US", "London, England, GB"],
  "detailLevel": "full",
  "forecastDays": 8,
  "maxHourlyItems": 48,
  "maxItems": 4,
  "maxConcurrency": 8,
  "proxyConfiguration": { "useApifyProxy": false }
}