This weather API turns a place name into typed daily rows — no API key, no separate geocoding step, no JSON spelunking: one row per day with min and max temperature, precipitation sum and probability, maximum wind speed, and the WMO weather code alongside a human-readable conditions label. Ask for up to 16 days of forecast and prepend up to 92 days of recent history in the same call.
Under the hood it is Open-Meteo — the open-data forecast service that has quietly become the developer default — fetched over our TLS tier with geocoding folded in and flattened into the same row envelope every other QuantumProxies collector uses. That is the honest pitch: weather as one more typed table in a pipeline that already pulls World Bank indicators and Google Trends curves, billed at $0.0002 per delivered day-row — and only for rows that actually arrive.
$0.0002 per delivered day, up to 30 days per run. Nothing delivered means nothing charged.
Send a location — any city or place name — and the collector geocodes it first, so "Rome" resolves to coordinates without you shipping a gazetteer. If you already hold exact coordinates, pass latitude and longitude instead and the geocoding step is skipped entirely.
Each day in the window becomes one flat row. Three details shape how you wire it in:
days sets the forecast horizon (1–16, default 7) and past_days prepends up to 92 days of recent history — which is what lets you join last month's weather onto last month's sales in a single run.weather_code for your logic and a readable weather label for your dashboard, so nobody maintains a code-to-text lookup table again.Everything is optional except a target: give location for a geocoded lookup or latitude/longitude for exact placement. days (1–16) sets the forecast window, past_days (0–92) reaches backward, and max_results caps the day-rows a run may return — 30 at most.
One row per day: date, temp_min_c and temp_max_c, precipitation_mm with precipitation_probability, wind_max_kmh, and the WMO weather_code with its weather label. The resolved location, country and the exact latitude/longitude used come back too, so you can verify the geocoder picked the Rome you meant.
$0.0002 per delivered day ($0.20 per 1,000). Nothing delivered means nothing charged, and the $2 monthly free credit covers roughly 10,000 days here. Volume tiers take up to 30% off.
$0.0002 per delivered day-row — $0.20 per 1,000. Concretely: a 7-day forecast for 200 cities refreshed every morning is 1,400 rows, or $0.28 a day. The free allowance is $2 of usage a month with no card, which here means 10,000 day-rows — enough to run a real dashboard before paying anything.
Billing counts delivered rows only. A run that geocodes nothing and returns nothing costs nothing.
Full honesty: Open-Meteo is free at the origin, keyless for non-commercial use, and excellent. If your whole project is one script that needs tomorrow's temperature, call it directly and pay nobody — including us.
What you are buying here is everything around the call. Open-Meteo wants coordinates, so you run a geocoding request first and handle its misses; the forecast comes back as parallel arrays keyed by variable name that you zip into records yourself; commercial use is meant to move onto their paid API tiers, which reintroduces a key to manage. This collector folds geocoding, zipping and typing into one request, returns day-rows in the same envelope as sixty-plus other collectors, and meters usage on the same key and invoice as the rest of your scraping — at $0.0002 a row, deliberately close to free.
A weekend version of this is easy: geocode, fetch, zip the arrays, done. What the weekend version grows into is the actual cost:
Here it is one POST, one schema, one bill. And if you would rather keep your own fetcher for everything, the residential proxies underneath are available on their own.
Use past_days to pull the weather that actually happened, join it onto sales or footfall by date, and let the model learn that umbrellas and rain correlate. The 92-day reach covers a full quarter's retraining window.
Route crews and deliveries around precipitation_probability and wind_max_kmh for the next one to sixteen days. One run per depot city, thirty rows or fewer each, fractions of a cent apiece.
Refresh destination pages with the coming week's weather labels and temperature bands. The readable label ships straight to the UI, so the pipeline is collector-to-CDN with no interpretation layer in between.
One run returns at most 30 day-rows (max_results), with the forecast horizon capped at 16 days and past_days reaching back up to 92 — size the window to the rows you need. Rows are daily aggregates by design: this collector does not return hourly curves, live radar or severe-weather alerts. Runs are independent, so a many-city sweep parallelises up to your plan's rate limit — 60 requests a minute on pay-as-you-go, 1,200 on the top tier.
Weather is open data: Open-Meteo publishes under open licences, so there is no terms-of-service cloud over reuse. The real caveat is accuracy, not legality — beyond about a week out, every forecast is an educated guess, ours included.
There is no key on either side: the upstream source is keyless and your only credential is the QuantumProxies API key you already hold. The free allowance is $2 of usage per month with no card, which at $0.0002 per day-row is 10,000 rows — a 7-day forecast for roughly 47 cities every day of the month.
The past_days parameter prepends up to 92 days of recent history to the forecast window, one aggregated row per day. It is designed for joining recent weather onto business data, not for deep climate archives — if you need 1950 onwards, that is a different tool.
No. Rows are daily aggregates: min and max temperature, precipitation sum and probability, peak wind and the WMO condition code with a readable label. Hourly resolution, radar tiles and severe-weather alerts are out of scope for this collector.
Twenty cents. Each delivered day-row is $0.0002, there is no subscription, and runs that deliver nothing are not billed. Volume tiers reduce the rate further, and GET /v1/scraper/collectors always shows the price your key actually pays.