Brolly — Privacy Policy

Last updated: 27 July 2026

Brolly has no accounts, no advertising, no analytics, and no trackers. Free installs store nothing that identifies you. If you buy the optional one-time upgrade, the payment library keeps the email address you paid with in your browser — that is the only personal information Brolly ever holds, and it is described in the table below. Every request the extension makes goes to a single server of ours; this page explains each of them, and exactly what each one contains.

What the extension stores, and where

Everything is stored in your browser, using Chrome's own extension storage. None of it is sent to us.

Data	Where	Why
The cities you chose — a home city, plus any extras you save — and your settings: units, badge appearance, clock format, notifications, Planner activities, which one-time hints you have dismissed, and which optional features are unlocked	`chrome.storage.sync`, mirrored to `chrome.storage.local`	So your settings survive a browser restart or a profile move. If you have Chrome Sync enabled, Chrome — not us — may sync this between your own devices.
The most recent weather data for each of those cities: the forecast, any official warnings, the next couple of hours of precipitation, and the temperatures behind the warmer-than-yesterday comparison. Also the planetary geomagnetic forecast, which is not tied to a city at all.	`chrome.storage.local`	So the toolbar badge and the popup are correct the instant you look at them, without waiting for the network.
The pressure mark you set by tapping the barometer	`chrome.storage.local`	So the dial can show which way pressure has moved since you set it. It is kept per city, on this device only, and is deliberately not synced.
If — and only if — you buy the optional one-time upgrade: the email address you paid with, and the fact that you paid, held by the payment library (ExtensionPay)	`chrome.storage.sync`	So that the features you bought stay unlocked, and so you can restore them on another computer by asking for a login link. Your card details never reach this extension or us — payment happens on Stripe's own checkout page. If you never buy anything, no email is ever stored.
The date this extension was installed, written by the payment library (ExtensionPay) that handles the optional one-time purchase	`chrome.storage.sync`	The library records it so that, if a paid tier is offered, it can tell a new install from an old one. It is a date and nothing else — no identifier, no address. It is written whether or not anything is for sale, and whether or not you ever buy anything.
Bookkeeping that keeps the rest working: which city the badge is currently following, which warnings you have already been notified about, and whether the last refresh for a city failed	`chrome.storage.local`	So a change of home city is noticed and refetched, rather than left showing the previous city's temperature; so a notification is not repeated; and so the popup can tell you when a check failed instead of showing an old reading as though it were current.

Nothing Brolly stores falls outside that table. Uninstalling the extension removes all of it, and no copy of it exists on any server of ours.

On precision. A saved city is held at the coordinates the place list gives for it, to four decimal places, and each stored item is filed under those same coordinates so that one city's data is told apart from another's. That is finer than the two decimals a request carries, and the difference is deliberate: these coordinates identify a town in a public place list rather than a position of yours, they are a label on your own disk, and the labels themselves are never sent anywhere. Every request rounds to two decimals first, as described below.

What we never ask for

Your device location. Brolly does not request the geolocation permission. You type a city name; that is the only location it knows.
Your browsing. It does not request access to the pages you visit, your history, your tabs, or your bookmarks, and it runs no code on any website.
Your identity. There is no account, sign-in, or identifier of any kind.

The requests it makes

To show a forecast, the extension asks our own server for the weather at the coordinates of the city you picked. That request contains:

A latitude and longitude, rounded to two decimals — about a one-kilometre grid square. This is deliberately imprecise: it is the centre of the town you chose, not a position, and rounding also lets everyone in the same town share one cached answer.
Your IP address, because every internet request necessarily includes one.
The publication time of the forecast already on your machine, once there is one — so the server can answer “that one is still current” rather than send it again. It is the weather model's own timestamp, the same for everyone holding that forecast, and it describes the data rather than you.

That is the entire payload of a forecast request. There is no identifier, no cookie, and nothing linking one request to another.

We keep no logs of it. The server stores cached forecasts keyed by rounded coordinates — never by IP, and never by user. It sets no cookies. The same request from a hundred people in one town is indistinguishable, because it is literally the same request. The one thing the server counts is how often its cache was hit or missed since it last restarted: running totals, held in memory, with nothing in them about who asked.

The extension makes four other kinds of request, all to that same server:

Official weather warnings — the same rounded coordinates. Where an official warning service can be asked about a point, our server asks it; for most of the world there is none to ask, and the request goes no further than our server.
Precipitation for the next couple of hours, where weather radar reaches — again the same rounded coordinates.
Geomagnetic activity, for the northern-lights forecast — no coordinates, and no parameters of any kind. What comes back is planetary: one set of figures that is the same everywhere on earth. So this is one fixed address, identical for every user. It is requested only if you turn that activity on and the city you chose is far enough north or south for an aurora to be possible there.
A city search — the term you typed, answered from a city list the server already holds. Search terms are not logged either.

Those answers are cached the way forecasts are: under the rounded coordinates, under the search term, or — for the planetary figures — under a single fixed address that carries nothing about you. Never under an IP address, and never under a user.

Who else is involved

Our server fetches forecasts from the Norwegian Meteorological Institute (MET Norway), passing along only the rounded coordinates. The short-range precipitation series comes from the same institute, requested the same way. So do official warnings, for coordinates in a broad band of northern Europe: that band decides only whether the request is worth making, and MET's warning service issues warnings for Norway, so a point elsewhere in the band is sent to it but has no warning to receive back. MET receives no information about you — not your IP, not anything else — because the request comes from our server, not your browser. Its data is used under CC BY 4.0.

For coordinates in the Americas, and in the US Pacific territories, official warnings come from the US National Weather Service (api.weather.gov). Our server sends it the same rounded coordinates and nothing else. As with MET, your IP address never reaches it, because the request is made by our server rather than by your browser. The service answers for the United States and its territories and refuses points outside them, which is how our server learns there is nothing to ask about a place and stops asking for a day. Its data is a US government work, in the public domain.

Geomagnetic activity, used only for the northern-lights forecast, comes from NOAA's Space Weather Prediction Center (services.swpc.noaa.gov). That request carries no coordinates and no parameters at all — it is one fixed address, and what it returns is the same everywhere on earth — so nothing about you or your city reaches it. It is named here for completeness, not because it receives anything.

The city list is built from GeoNames, also used under CC BY 4.0. It is stored on our server; no request is made to GeoNames when you search.

The server runs on Cloudflare Workers. Cloudflare processes the request as our infrastructure provider and may retain transient network-level logs under their own privacy policy.

Notifications

If you leave severe-condition alerts on, notifications are generated on your own machine from the forecast and the official warnings already stored there. Nothing is sent anywhere to produce them, and no push service is involved.

Children

Brolly is not directed at children. Free installs collect no personal information from anyone, of any age. The only personal information that can ever appear — a buyer's email, held in the browser by the payment library — comes from an optional purchase a parent or guardian would have to make.

Changes

If this policy ever changes in a way that affects what leaves your browser, the change will be described here with a new date. The dated notes below are the record.

27 July 2026 — Brolly gained official weather warnings, short-range precipitation and a northern-lights forecast. Each is a further request to the same server, carrying the same rounded coordinates as a forecast or, for the northern lights, no parameters at all. Two upstream services are named here for the first time: the US National Weather Service, which receives those rounded coordinates from our server, and NOAA's Space Weather Prediction Center, which receives nothing. The same day, the optional paid upgrade began storing the buyer's email via ExtensionPay; the opening paragraphs were rewritten so they no longer claimed Brolly holds no personal information at all.

Contact

Questions or concerns: andreas.nelvik.engebretsen@gmail.com.