No. RetireSafe is an analytical tool. It simulates thousands of futures from the inputs you provide and tells you what the engine computes. Decisions about your money should involve a licensed advisor and consideration of factors no calculator can model.

Stocks, bonds, cash and inflation are drawn together as one correlated set, so a bad year for one is a year the others feel. Lifespan is drawn from a curve set by country, sex and health (the Gompertz law), and the chance long-term care begins rises with age. What comes out is a spread of outcomes, not a forecast. The value is in the shape of that spread rather than any single number. If the world departs substantially from its historical behaviour, this engine will be miscalibrated alongside every other model built on historical data.

Your plan is calculated on our servers in the EU, not on your device. Its inputs are used for that calculation and not kept, and the results are deleted within the hour of reaching your browser. If you create an account and explicitly save a plan, its inputs and results are written to servers in the EU, where only your account can read them. There is one exception you did not ask for: if a simulation run fails, that run's inputs travel with the error report so we can reproduce the failure and fix it. The privacy policy sets out exactly what that covers.

Free is single-person, around 10 inputs, long-term care at your country's typical cost, and the headline answer with the full range of outcomes charted. Sign in and you keep 3 saved plans you can re-open from any device. Premium adds couples, your home, your own care costs with family help, spending split into seven categories, how your spending shifts with age, what's most likely to break your plan, every change that could fix it, the best way to take your money out, choosing which account you draw down first, choosing how each account pays out, the highest spending you can sustain, what your tax bill is made of, what tax costs you each year, which account each year's withdrawals came out of, every chart's numbers as a year-by-year table you can download, whether moving money into your Roth pays off (US) and a PDF report. Full table on the pricing page.

The methodology is the product. Engine improvements ship as code; they don't generate a recurring billing event. We bill once because that's what the work costs, not because SaaS conventions say to charge monthly.

Premium-only. Both partners get their own lifespan, and the two are linked rather than drawn independently, because partners share a household, habits and health shocks. The strength of that link is empirically calibrated (a Frank copula over two Gompertz death-time draws, α≈3.4 by default). On the first death, each spending category is scaled separately (essentials fall by less than half, housing barely moves), pension survivor shares kick in, and the simulation continues for the surviving partner.

Twenty, listed in full on the country page. Each one carries its own researched numbers rather than a shared default: when the state pension starts and how it rises, what a surviving partner keeps, what healthcare and long-term care cost, what a retired household spends, and how your withdrawals are taxed. Every figure shows where it came from and the date it was read. Pick your country on the first step and your plan starts from its numbers.

None, unless you ask for one. By default the engine spends exactly the plan you enter, every year, rising with prices, and tells you how often your savings pay it. On Premium you can let your spending follow the markets on the spending step: you pick a rule that caps what you take out each year (a fixed amount starting at 4% of your savings, a fixed share of what is left, or one that follows your balance and age) and how far essentials and extras may be cut when it binds. Either way, your results show how much of your plan you actually got.

On every plan, Free included. The chance long-term care begins rises with age. In any future where it does begin, the plan pays a year of care on top of your other spending, less the everyday living costs that care replaces, and that cost grows 1.8% a year faster than prices in general. Free plans use your country's typical cost of care. Premium lets you set the yearly cost yourself and say how likely family is to take on part of the care.

Premium-only. After the main 10,000-future run, the engine varies each of your inputs across its plausible range and measures how much each one swings your chance of the money lasting. The result is a ranking of what is most likely to move the needle for your specific plan, not a generic list. Formally these are Sobol total-effect indices.

The engine picks changes you could realistically make and re-runs the whole simulation for each: saving 500 more a month, retiring 2 years later, putting 15 percentage points more of your savings into shares (more risk, not less), paying for care at family-care rates, and keeping your home where the plan would sell it. Another set keeps your plan and changes how you take the money out: which account you draw from first, your planned conversions, and how each account pays out. You see how each one shifts your chance of the money lasting and what's left at the end. A free run shows one of them, a different one each run; Premium shows every one that applies.

No. You can fill everything in and see your results without signing in, and that run stays saved in your own browser. Keeping plans that you can re-open on another device needs an account, which signs you in by emailing you a one-time link. There is no password to remember.

Tell us on the contact page. Include the engine version, which is in the footer of your results, and the inputs you used. Every saved scenario records the engine version that produced it, so we can reproduce exactly what you saw.