the 3am question

Will your money last — and what changes it if it won't.

Put in what you have, what you earn, what you spend and when you want to stop. FIRE Model runs that plan through thousands of possible futures — including the bad ones — and gives you three things: the odds it holds, the year it breaks when it breaks, and the shortest list of changes that fix it.

An answer you can defend to yourself at 3am, instead of a rule of thumb you read on a forum.

One payment, yours for life. No subscription required — or rent it for $63 a week if you only need it this month.

The nearest thing to this costs $129 a year.

Two years of a subscription costs more than owning this outright, and you can check that in one click. That is the whole price argument; it does not need a comparison table with ticks in it.

Planning is episodic. You do it hard for a month, then revisit it once a year. Paying every month for the eleven you are not using is a shape that suits the seller.

Lifetime

$249

Paid once. Every future version. Nothing to renew.


Or rent

$63 a week

For the month you actually need it. Five weeks of rental costs more than buying it outright — if you think you will come back to this next year, buy it.

Get FIRE Model

Four numbers, and the reason behind each one.

Does it hold?

A success rate that always arrives with its definition, its sample count and its margin of error — never a bare gauge.

When does it break?

A plan that runs dry at 94 is a different plan from one that runs dry at 68. Everything else in this category collapses both into one percentage.

What fixes it?

One change at a time, ranked by how much it moves. Spending less, stopping later, paying lower fees — measured, not guessed.

When could you stop?

The earliest year that still clears your target, and the highest spending that does. Both from the same engine as everything above.

the useful bit

Not “you're at 87%”. Rather: spending $400 a month less takes you to 96%; retiring a year later only gets you to 89%; dropping your fund fee by 0.4% is worth more than both.

Nobody changes their life because of a percentage. They change it because they can see which lever is short.

What it models, and what it can't.

This list is inside the app too, printed next to your results — not tucked into an FAQ where it counts as having been said.

Modelled

  • Sequence-of-returns risk — the thing that actually ruins retirements
  • Inflation, per simulated year, with everything shown in today's money
  • Fees, as an annual drag, because they are the most reliable predictor in the whole model
  • Debts, amortised monthly to get the annual figure right
  • Guaranteed income that does or does not rise with inflation
  • Spending split into what you would not cut and what you would

Not modelled

  • Mortality. The plan runs to a fixed age. Modelling death makes the success rate go up, and a number that improves when you get sicker is not one we will print.
  • Full national tax engines. The law changes every year and a lifetime licence holding last year's rules is worse than an approximation you can correct.
  • Live prices, bank connections, or your actual portfolio. This is a planner, not a tracker.
  • Long-term care, currency risk, or a plan in two currencies at once.
  • Anything AI. There is no chat, no “explain my plan”, and no language model anywhere near your retirement.
  • Your own behaviour — and the most common way a real retirement fails is a person selling at the bottom.

About “97years of market data”

The historical engine draws on 19282024of US returns and inflation. That is one country's record, and it belongs to the most successful equity market of the twentieth century — which is survivorship bias, and we say so on the results screen rather than treating the number as a trust badge. A second engine exists precisely so you can assume less and see what it costs.

How it works, for the people who ask.

It runs on your machine

The maths, the market data and your plan all live in your browser, and your plan stays there — no salary, balance or date of yours is stored on a server. The app talks to us for one thing only, to confirm you own your licence, and it keeps working when it can't reach us. Every page but the app itself is served a policy that cannot open an outbound connection at all.

The same number, every time

Each plan carries a seed. The same plan, the same seed and the same engine version give the same answer on any machine, forever. A planner whose number quietly moves overnight is not a planner.

It refuses to guess

No spending figure, no horizon, an allocation that doesn't add to 100 — it says so and stops. A confident percentage built on our defaults would look exactly like one built on your numbers.

Three free calculators, same engine.

No account, no email. The reason free calculators in this category all disagree with each other is that no two share assumptions — these three share theirs with the full app.

Questions worth asking before you pay $249.

Is this financial advice?

No, and that is a set of rules rather than a footer. The app never recommends, ranks, selects or names an investment — no fund, no ticker, no “recommended allocation”. It shows what happens to your assumptions when you change them. There are no affiliate links, no broker referrals and no money from anyone who sells a financial product, which is what makes the first part credible rather than decorative.

What does a 95% success rate actually mean?

That in 95% of the sampled sequences, under this model and your assumptions, the plan funded its floor spending every year to the end. It is not a 95% chance you will be fine. Every time the number appears it carries that definition, its sample count and its margin of sampling error — because 94% and 96% are the same answer and printing them as different ones is dishonest.

Where does my data go?

Nowhere. Your salary, your balances and your plan are stored in your browser and stay there — there is no account holding a copy and no switch that would start one. Every page except the app itself is served a policy that cannot open an outbound connection at all, and you can check that with curl.

Does my plan sync between my laptop and my desktop?

Not in this version, and we would rather say so on the pricing page than behind a switch. Your plan lives on the device you build it on. You can export the whole thing as JSON at any time — stamped with the seed and engine version, so the numbers can be reproduced from it — but this version cannot read one back in, so treat it as your copy rather than as a way to move house. Syncing your own devices is the next thing we want to build here, and it will be included in the price you already paid.

Does it work offline?

Yes. The market data ships inside the app, so once it has loaded once there is nothing to fetch. Open it on a plane a year from now and it works.

Which countries does it handle?

Any — the engine works in tax treatments (taxable, tax-deferred, tax-free, cash) rather than in national account types, and labels them for your country. Tax itself is an editable rate table with starter values for the US, UK and India, each carrying the date it was written. It is an approximation you can see and correct, not a filing.

Why one payment instead of a subscription?

Because planning is episodic. You do this hard for a month and then revisit it once a year, and a subscription is a bad shape for that. A weekly rental exists for the month you actually need it — and the buy page will tell you when renting costs more than owning, because it usually does.

Is there a Mac app?

Not in this release. It runs in the browser and installs as an app from there.

Buy it once. Ask it again in ten years.

$249, paid one time. No subscription required. Your plan stays on the device you build it on — this version does not sync between machines — and nobody ever tries to sell you anything inside it.

FIRE Model is a LifeTime app. It is a modelling tool, not financial advice, and it takes no money from anyone who sells a financial product. Have a professional confirm anything you plan to act on.