Ask about the weather in plain language and get a grounded, specific answer — across all your saved locations. Not a generic chatbot guessing, but an assistant that reads your actual forecast and answers from it.
The problem
General-purpose chatbots will happily “answer” a weather question by making something up — they have no live data. And reading a raw forecast yourself means scanning tables for the one detail that matters: do I need an umbrella for my walk at 5pm?
How it works
The assistant is grounded on your real forecast bundle — current conditions, hourly and daily data, air quality and official alerts for the locations you’ve saved. Your question goes to the model together with that data, so the answer comes from the forecast, not the model’s imagination. Ask about any of your locations by name; it resolves which one you mean and answers from there.
What makes it genuinely useful
- Forecast-delta narration. It can tell you what changed since yesterday’s forecast — “Saturday’s rain moved six hours earlier” — not just today’s numbers.
- Honest about uncertainty. When the models disagree, it says so, instead of inventing false confidence.
- Rule authoring in words. Describe what you want to be warned about in plain language and turn it into a weather-monitoring rule — the assistant is a natural front door to the watchdog.
Why it’s different
- Grounded, not guessing. Answers are backed by your live forecast data.
- Multi-location. One question, all your places.
- Safeguarded. It won’t fabricate forecasts or alerts; it reads them.
Pricing
The AI assistant is a Pro feature and runs on AI credits. Every plan includes welcome credits so you can try it; Pro includes a monthly allowance, and credit packs top up any time.
FAQ
Is it just ChatGPT for weather? No — it’s grounded on your actual forecast data, so it answers from real numbers rather than guessing.
Can it answer for multiple cities? Yes — ask about any saved location by name.
What are credits? AI features consume credits; plans include a monthly allowance and you can buy one-time packs.
Will it make up a forecast? No — it’s constrained to the forecast data it’s given, and flags uncertainty instead of inventing precision.