Who builds this, and how

Who runs this

That's all of it. One person writes these tools, and the company above is the legal entity he operates under. There is no team and no newsroom behind the calculators. We won't pad this page with an invented team, an invented founding date or an invented user count. The point of the page is that you can check what it says, and on a page about money a claim you can't check is worth nothing.

The CNPJ is a public Brazilian company register number. Publishing it exposes nothing that isn't already searchable, and it puts a legally identifiable entity behind every number on this network. The same identification is on the privacy policy and the contact page, because data protection law wants to know who is responsible. A way to send a message doesn't answer that.

Why I started building these

My name is Douglas Santos. Most of what's on this site began as something I needed myself. That history is the reason the tools are shaped the way they are, so it's worth a few paragraphs.

The pattern never changed. I'd have a concrete question, usually how much material a job really takes, or what a fee costs once everything is counted. Then I'd go looking for a calculator. There's never a shortage of calculators. What I kept running into was that the one I found answered a slightly different question from mine, and didn't say so.

Some were plainly wrong: a marketplace fee figured on the item price alone, when the fee is charged on the shipping too, so the number came out low every time. Others were right and still useless. They printed one confident figure for a quantity that has a wide range around it and no honest single value. A few wouldn't admit that a question has two constraints and answered only the easier one. And plenty wanted an account or an email address, or put two lines of arithmetic behind a queue of interstitials.

So I wrote my own, one at a time, as each need came up. I had the problem before I had the tool, which is why they tend to handle the awkward part instead of skipping it. The fee comes off the full buyer total. The second constraint gets checked even when the first one already produced an answer.

None of this was planned as a network. It piled up. At some point there were enough of them that keeping them on my own machine had stopped making sense, and the honest move was to put them where someone else with the same question could find them.

They're all free, and open in the way that matters when you're trusting a tool with a number: no account, no signup, no usage limit, no paywall and no watermark. The working is shown on the page instead of hidden behind the result, and there's an embed if you want one on your own site. Most of them never send your input anywhere at all. The arithmetic runs in your browser, and you can confirm that in the network tab. Advertising pays for it, and nothing else does.

The second failure does more damage, because it survives review: the tool hides its own limits. It prints a number to two decimal places when the model underneath has a spread of forty percent. It puts a green badge next to a result nothing can justify. It answers a question that needs a professional, then buries the disclaimer in gray type at the bottom of the page, below the ads, collapsed behind a chevron.

A number carries authority just by being a number. Producing one and keeping quiet about what it can't tell you is the failure this network was built to avoid.

The rest of this page is the method: how a tool gets built, how the arithmetic is checked, what the tools refuse to do, and where I have been wrong so far.

How a tool is actually built

One rule, and the file layout enforces it: the math lives in its own module, outside the page. Thirty-one of the live tools have one. It's a plain JavaScript file of pure functions, no DOM, numbers in and numbers out, sitting next to the interface that uses it.

This isn't tidiness for its own sake. When the arithmetic is buried in an inline script, the only way to check it is to open the page and squint at the result, and that is how a wrong calculator stays wrong for years. When the math is a module, a script can call it directly and compare the output against a value that came from somewhere else.

That script is scripts/verify-math.mjs. It runs 35 checks against the live tools today, and all of them pass. The count matters less than the rule the file follows: an expected value derived by re-running the same formula proves nothing. So every case in it has an answer from outside the code, either an internationally agreed definition, an algebraic identity, or a calculation done by hand.

In practice that means checking that an inch is exactly 0.0254 meters and a pound is exactly 0.45359237 kilograms, not the rounded values that circulate, and that one year at a 4.5% APY yields exactly 4.5% at every compounding frequency, since that is what APY means.

Content has a gate of its own. A build script blocks the release of any page it audits that comes in under 1,200 words, five sections or four answered questions, so an explanation can't quietly shrink to a heading and a form. Thirty-seven pages are inside that gate. The rest are marked exempt in the site register, one by one and in writing, because for those pages a long explanation would be work that never turns into a reader.

The tools tell you when not to use them

This is the part we'd most like to be judged on, because it's the part that costs something. Each of these is a decision in the source code. None of it is a marketing line.

The wire size calculator refuses to answer. Cable sizing is two independent problems. Current-carrying capacity protects the cable, voltage drop protects the equipment, and the answer is the larger of the two. If no size in the table satisfies one of them, the tool returns a failure and no size at all. Handing back the winner of the criterion that did have a solution would mean printing a gauge that melts under the load, dressed up as a calculated answer.

The severance calculator says in the result itself that the binding figure is the one on the signed settlement document, not the one on our page, and lists what can move it.

What we do not do

• We don't sell your data. There is nothing to sell: for ordinary browsing we store a salted hash instead of an IP address, which is why the privacy policy admits we usually can't find records tied to you even when you ask.
• We don't ask you to sign up. None of the tools has an account, a login or an email field.
• Nothing here sits behind a paywall or a watermark. Advertising pays for the network, and the cookie policy covers that in full.

Where we were wrong

Every one of these shipped or nearly shipped, and each one is documented in the source where it was fixed. We publish them because a page that claims to be careful is worth nothing without examples of the carelessness it caught.

One whole category of error never reaches a user and still deserves a name: the off-by-one. Dividing a length by a spacing gives you the number of gaps, not the number of posts, joists or studs. The missing one is always the one at the edge, which is where the load has no support. Several tools in the construction group carry an explicit comment at that line for exactly that reason.

How to correct us

Write to [email protected]. A report that a calculator is wrong is the most useful message we get, and fixes to the arithmetic are the only work here that jumps the queue. A wrong calculator is worse than no calculator.

For us to act on a report, it needs the tool, the exact values you entered, the result you got, the result you expected, and ideally the source that says otherwise: a formula, a code section, a published table. "It is wrong" can't be acted on. "At a cost of $80 and a price of $100 it shows a 25% margin, and the margin is 20%" gets fixed.

If we got something wrong here, we'd rather hear it from you than never find out. The terms of use set out which tools need which kind of professional, tool by tool, and the contact page covers privacy requests, security disclosure and everything else.