Why Every Calculator Here Runs in Your Browser

Why a Calculator Should Not Need a Server

There is no technical reason for a mortgage calculator to send your salary to anybody. The arithmetic is a handful of multiplications, your browser can do several billion of those a second, and the round trip to a server is slower than doing the work locally.

Yet a great many calculators do exactly that, posting inputs to an endpoint and rendering whatever comes back. Sometimes it is architectural habit. Sometimes the inputs are the product, and the calculator is a form collecting income, debts, health metrics or business figures under the impression of a utility.

Every tool on this site computes in the page. The mortgage calculator never learns your income; the BMI calculator never learns your weight; the net worth calculator never learns anything at all. Not because we promise not to look, but because the figures never leave the machine you typed them on.

You can verify this rather than trusting it. Open your browser's developer tools, switch to the network tab, and use any calculator on the site. Nothing will be sent. Better still, load a page, disconnect from the internet entirely, and keep using it — every calculator continues to work, which is only possible because there was never a server involved.

The Search Box Makes No Requests Either

A site search is the usual exception even on privacy-conscious sites. Search-as-you-type normally means a request per keystroke, and those requests form a rather intimate record — what someone was looking for, in what order, and what they abandoned halfway through typing.

The index for all one hundred calculators here — names, descriptions, categories and keywords — is small enough to compile directly into the JavaScript that ships with every page. Searching filters an array already in memory. There is no autocomplete endpoint because there is no endpoint.

This costs a few kilobytes on every page load and buys the property that search behaviour cannot be logged, because it never arrives anywhere capable of logging it. It is a reasonable trade at this scale, and the reason to mention it is that the alternative is invisible to users who have no way of knowing which design they are using.

Consent Before Scripts, Not After

This site carries advertising, which is how it stays free, and advertising involves third-party JavaScript that we do not control.

The arrangement here is that the advertising loader is not injected into the page until consent is granted. Not loaded-then-disabled, not loaded-with-tracking-off — not present in the document at all. Decline, and the tag is never added, so it never runs, never sets anything and never makes a request.

This is the distinction that a great many consent banners elide. A banner appearing above scripts that have already loaded is a notification, not a choice. The order of operations is the entire substance of the thing, and it is visible in the network tab of any browser: decline, and there is nothing third-party to see.

We set no cookies of our own. Your theme preference and your consent decision are kept in local storage on your own device, which is not transmitted with requests the way a cookie is, and both can be cleared by clearing site data.

What This Design Cannot Do

It is worth being straight about the limits, because privacy claims are cheap and the interesting part is always what is excluded.

Client-side computation means we cannot save your work. There are no accounts, no saved scenarios and no history, because all three require a server holding your data. The share button on each calculator encodes your inputs into the URL instead — that link is generated in your browser and is not recorded anywhere, and it is worth remembering that anyone you send it to can read the figures in it.

It also means we cannot personalise anything, cannot show you what you used last week, and cannot email you results. Those are real conveniences and they are genuinely absent. The trade seemed obviously right for tools that handle salaries, debts, health metrics and business figures.

And it does not make us the only party involved. Your browser, your network, your internet provider and the advertising you consented to all exist independently of anything we do. What we can say precisely is narrower and more checkable than a broad promise: the numbers you type into the budget calculator, the retirement calculator or any other tool here are never transmitted, because the code that would transmit them does not exist.

What We Actually Know About You

Claims about privacy are most useful when they are specific, so here is the complete list of what a visit to this site produces.

Your browser requests HTML, CSS, JavaScript and font files, and those requests reach our host. That means an IP address, a user agent string and a page path exist in server logs, as they do for every website that has ever been served over HTTP. It is not nothing, and it is the irreducible minimum for delivering a page at all.

Beyond that: no cookies of our own, no analytics beacon fired on page load, no form submissions except the contact form when you deliberately send one, and no request of any kind triggered by using a calculator. The last point is the checkable one — open the network tab, use any tool on the site, and watch nothing happen.

If you consent to advertising, third-party scripts load and they have their own data practices which are theirs rather than ours. That is the honest boundary of what we can say. We control whether those scripts are present; we do not control what they do once they are, which is why the consent decision is the meaningful one and everything downstream of it is not our promise to make.

The Trade-Offs This Design Forces

Client-side computation is not a free win. It rules out several things that users legitimately want, and pretending otherwise would be the same overclaiming this article is arguing against.

Nothing persists. No accounts, no saved scenarios, no history of what you calculated last month, because every one of those requires storing your figures somewhere we can reach. The share link is the substitute: it encodes your inputs into a URL generated entirely in your browser, which you can bookmark or send. It is also worth remembering that anyone receiving that link can read every figure in it.

Nothing is personalised. The site cannot show you your most-used tools, cannot remember your preferred currency across visits beyond what local storage holds on your own device, and cannot email you a result. These are real conveniences and they are genuinely absent.

Every page carries its own weight. Compiling the search index into the JavaScript costs kilobytes on every load rather than fetching them once when someone searches. It is a deliberate trade of bandwidth for the property that search behaviour cannot be logged.

Whether that set of trades is right depends on what a tool handles. For a recipe converter it would be over-engineering. For calculators that take salaries, debts, health metrics and business figures, it seemed clearly worth the cost — and the fact that you can verify it in thirty seconds with developer tools open is the part that makes it a design rather than a promise.