Kessavo
Current Engine:v3.0.2 (2026 IRS/SSA Rules)

What's new

Kessavo is in active development. Here is what has shipped and what is coming next — including improvements driven directly by beta tester feedback.

v3.0.2 — August 2026

Latest Release

v3.0.1 — August 2026

v3.0.0 — August 2026

Earlier releases — v1.0.0 through v2.4.2Show

The releases that built the current model, from the first public beta through v2.4.2. Kept for anyone checking when a rule or a figure changed. Descriptions are the ones written at the time.

v2.4.2 — August 2026

  • Fixed (Amounts You Typed Were Not The Amounts Used): Money and percentage boxes reformatted themselves as you typed, which put the cursor after the “/ mo” or the “%” and threw the next digit away. Typing 2,500 into a monthly income could store 2. Typing a plain 0 emptied the box, a decimal point vanished as you typed it, and a comma in 10,700 cleared the whole field. Boxes now leave what you type alone until you move on, then tidy it up — and they accept commas and dollar signs.
  • Fixed (A Debt Payoff Year With No Debt): “Debt payoff year” sat on a different step from the monthly debt payment it refers to, and that payment was tucked inside a collapsed panel — so most people were asked when a payment stops without ever being asked for a payment. The payoff year now sits directly under the amount and only appears once you have entered one. Printed reports no longer show a payoff year above a $0 payment.
  • Fixed (Help Text Was Being Cut Off): Explanations under some fields were trimmed mid-sentence, including the one noting that employer contributions always stay tax-deferred. Every field is now labelled, then the box, then its explanation, so the explanation can be as long as it needs to be and the boxes still line up across the page.
  • Added (Why A “One Change” Row Can Be Negative): “Claiming Social Security at 70” can show a negative figure. It is not an error: deferring pays nothing during the years you defer, so the model covers that spending from savings, and money withdrawn early stops compounding. The larger benefit afterwards narrows the gap but does not always close it by age 100 at a high assumed growth rate. The panel now says this, and states plainly that the balance left at 100 is one measure — it does not price a benefit that is inflation-adjusted income for life.
  • Improved (Clearer Saving Questions): “Saving rate” now says it is the share of that salary going into a 401(k), 403(b), TSP or similar, and notes that employer money is entered separately. “Saving directed to Roth” now says it means the share of the money you contribute, not your existing balances.
  • Improved (Calculator Layout And Spacing): Input boxes now line up across each row, spacing between fields follows one scale rather than ad-hoc values, and the “Start your plan” button no longer appears while you are already several steps into a plan.

v2.4.1 — August 2026

  • Fixed (Report Pages Were Overlapping): The first page of the printed report had its headline, chart caption, figures and closing paragraphs drawn on top of one another, and the calculation trace stretched two rows across a whole page. The trace now has its own page and every page was checked line by line.
  • Improved (A Real Balance Chart In Print): “Balance over time” is now drawn properly for print instead of being photographed from the screen. It has readable money and age axes, gridlines, your selected claiming age drawn over the others for comparison, and it stays sharp at any zoom. It also no longer disappears if you switch tabs while the report is being prepared.
  • Fixed (Range Wording When Nothing Runs Out): When every tested market path lasted the full horizon, the report said your savings could last “anywhere from age 100 or beyond to age 100 or beyond”. It now says plainly that no sampled scenario ran out, and that what separates them is how much is left at the end.
  • Added (Download Your Report): The results page can produce a print copy of your plan again. Report export had been unreachable since v2.3.1-beta8 — the control lived in a sidebar that was replaced by the step navigator, so there was no way to produce one at all.
  • Added (Your Changes Reach The Printed Copy): If you change your spending or a claiming age on the results page and then download the report, the report is built from what you were looking at. Your original scenario travels with it, and every figure that moved shows both values. Download without changing anything and none of that appears.
  • Fixed (One Model Behind Both Reports): The results page and the report export computed the same figures separately, and had drifted apart. For a household already receiving Social Security, the page ran its 300-path stress test at the claiming age in the input field while the headline beside it came from the age the household actually claimed — two different scenarios in one report. Both now come from a single calculation.
  • Fixed (Report Download Could Hang): Switching to another browser tab while a report was generating stopped it permanently, leaving the button on “Preparing report…” with no error. A chart that cannot be drawn now also leaves its written explanation in place rather than failing the whole download.
  • Improved (Order Of Returns, Shown Side By Side): The same set of returns run best-first and worst-first now appear as a comparison rather than a single difference figure. Same returns, same average — one ordering can fund the plan to 100 while the other runs out decades earlier, which is the clearest thing the model can say about risk.
  • Improved (What Moves The Answer Most): The list of single changes and their effects is now ranked by the model, re-running the whole projection once per change, instead of showing a fixed pair. Each row names the exact input movement so you can check the model did what you expected.
  • Improved (The Headline Says Which Path It Is): The primary outcome is labelled “steady-return path” and now reconciles itself with the funded-paths figure beside it. That headline comes from a projection earning the expected return exactly, every year, which makes it optimistic against the varied runs — so “Funded through age 100” could sit next to “52% of modeled paths funded” with nothing explaining the gap. The gap is year-to-year variation, and the page now says so.
  • Fixed (One Answer About Social Security): The income step called Social Security “required for the model”, the person card badged it “needs input”, and the status strip said “ready for results” — three answers to the same question, none of which blocked anything. It is now a single advisory state: the strip reads “Ready — 1 thing worth adding”, the section says “needed for a realistic result”, and the results page carries a note when a plan is modeled with no Social Security in it. Plans genuinely without it still run, unchanged.
  • Fixed (Phantom Saved Plan): Opening the calculator and leaving without entering anything no longer saves a blank plan. Previously the next visit opened with “Picked up where you left off” over a scenario reading “Born 0 · retire 0 · $0 starting savings”. Entering a single figure still saves and still resumes as before.
  • Fixed (One Name Per Growth Rate): The retirement account growth rate is now called that on the input as well as in the PDF. The Swiss input had it as “Expected annual return” while the report said “Retirement account growth rate” — one field under two names, which v2.3.1-beta1 had already corrected once. The check that guards it now covers both surfaces instead of only the report.
  • Improved (Outcome Bar Says What It Measures): The bar under the headline result states its own quantity and value — ending assets as a share of starting savings — rather than only saying what it is not. It sits directly beneath the “Modeled paths funded” percentage, which it has never represented, and the two are now told apart on sight and by screen reader.
  • Fixed (Single-Person Wording And Layout): The income step no longer describes a one-person plan as “income from both people”, no longer labels the only person “Person 1”, and no longer carries prototype instructions about testing the two-person layout. The results page no longer draws an empty panel where a spouse would go.
  • Improved (Visible Engine Version): The calculator header now shows the full engine version and tax-rule year directly under “Retirement planning, made inspectable.”, so the version you are running is visible while you enter a scenario rather than only on the results page, sidebar, PDF and changelog.
  • Improved (Results Chart Hierarchy): The duplicate report balance chart has been removed, and the Retirement cash flow chart now uses the full projection width. The interactive what-if section still retains its balance comparison for testing changes to a scenario.
  • Improved (Single-Person Clarity): The calculator now describes a single-person model accurately instead of referring to two separated people, and the first-step “Build the timeline first” title stays on one line in the Swiss layout.
  • Improved (Clearer Calculator Identity): The production calculator header now uses “Retirement planning, made inspectable.” as its single identity line instead of repeating product and prototype labels.
  • Fixed (Result Scenario Identity): Results now distinguish single-person and two-person plans from the submitted household mode, including single-person outcome wording and timeline events. The spouse-dies-first option is also named as a one-direction scenario so its scope is clear.
  • Improved (Results Charts): The results story now shows a data-backed balance chart with two-year points, an explicit terminal balance, retirement and survivor markers, a cash-flow chart, and a Monte Carlo balance-range chart with a table alternative. Calculation-trace typography was reduced for easier scanning, and the redundant inner calculator brand row was removed from the Swiss input surface.
  • Fixed (Swiss Prototype Fidelity): The production calculator now renders the approved Swiss Ledger input prototype itself. Production-only fields — state tax, survivor view, earned income, other regular income, one-time amounts, military pensions, and Roth conversions — are restored inside that same aligned layout instead of leaving the older step components behind the Swiss shell.
  • Improved (Swiss Ledger Calculator Flow): The approved Swiss Ledger direction is now the production calculator experience. The sidebar is replaced by a top progress navigator, the complete spouse-aware input flow remains intact, and the results page presents the scenario story with immediate what-if recalculation for spending and Social Security claiming ages.
  • Improved (Accessible Calculator Controls): Shared calculator fields now connect every visible label to its input, expose helper text to assistive technology, and keep toggle labels from double-activating their switches. The real calculator flow was checked in Chromium and WebKit, including couple, survivor, in-payment, pre-1960, and mobile scenarios.
  • Fixed (Age-100 Results Agree Everywhere): The balance chart and year-by-year tables now use the same final modeled balance as the results headline and PDF at age 100. The Monte Carlo range and sequence-risk charts use that same endpoint too, so the last plotted value no longer represents the start of age 100 while the summary represents the end of it.
  • Fixed (Full Retirement Age By Birth Year): Social Security claiming calculations now use SSA's graduated full-retirement-age schedule for cohorts born before 1960 instead of applying age 67 to everyone. The results page and PDF also include a shared calculation trace showing how submitted balances, timeline, income, withdrawals, endpoint, and range test connect. The calculator collects birth year and whole-year claim ages, so month/day-specific exceptions remain outside the model.
  • Improved (Results Explain The Number): The results story now separates Social Security amounts entered as reference figures from the household amount actually modeled after claiming and spousal rules. It also names the taxable/cash growth and cost-basis assumptions, explains the spending additions that sit above the everyday monthly figure, and labels future-dollar spending separately from today's-dollar balances.
  • Improved (Less Probability-Like Language): Monte Carlo output now says "Modeled paths funded" rather than "Success rate," and the verdict's percentage is labelled as input completeness rather than a confidence score. These are model descriptions, not probabilities or recommendations.
  • Fixed (Exact-Age Depletion Boundary): A projection that depletes at the selected reference age is no longer described as lasting beyond it. The online and PDF report wording now keeps the model's selected path and horizon meaning explicit.
  • Fixed (Unsafe Input Parsing): Formatted dollar and percentage fields now preserve a typed negative sign instead of silently turning a negative value into a positive one. Values are normalized at the calculator boundary before they reach the model, including saved plans and nested income entries.
  • Fixed (Stale Household Overrides): Changing from a custom "Other" state to a supported state can no longer leave the custom no-tax override active behind the scenes. Survivor modeling is also cleared when couple mode is turned off, and impossible household year or survivor-age combinations are stopped before results.
  • Fixed (Workflow State): Clearing saved entries no longer recreates a blank saved plan after the wizard resets. Invalid household timelines restored from browser storage are sent back to the household step instead of being allowed through to results.
  • Fixed (Estimate Readiness): The running estimate now waits for at least one balance, income, or savings input before showing a solvency result, so dates alone cannot produce a depletion warning. An untouched Social Security estimate also clears when its salary source is removed, while user-entered values remain intact.
  • Fixed (Two Fields Called The Same Thing): On the savings step, opening "Refine these" showed two percent boxes side by side, both labelled "Growth rate". One sets how your taxable and cash balance grows; the other sets how your retirement accounts grow, and is the rate every projection and simulation in your report is built on. Nothing on screen said which was which, and on a phone they stack with no clue at all — so it was entirely possible to change the wrong one, see your result barely move, and reasonably conclude the calculator was broken. They are now "Taxable / cash growth rate" and "Retirement account growth rate", each saying what it applies to.
  • Fixed (Your Report Now Uses The Same Names You Do): The same two figures appeared in your PDF report as "Cash yield" and "Expected nominal return" — three names for two fields across two places, so a number you entered came back to you under a name you had never seen. Your report now uses exactly the labels the calculator uses. "Cash yield" was also the wrong word for what the model does: it grows your cash balance at that rate rather than paying it out as income, and those are taxed differently.

v2.3.0-beta1 — August 2026

  • Added (Sequence Of Returns): Your results page now shows the same set of yearly returns run twice — once with the good years first, once with the bad years first. Same money, same years, the same average return in both. The only difference is the order they arrive in, and for most plans that alone moves the ending balance by a very large amount. The 300-run simulation already accounted for this, but only as a percentage; this is the chart that shows you why the percentage is what it is. Neither ordering is a forecast — they are the two extremes, shown together because a single average growth rate cannot express order at all.
  • Added (If You Changed One Thing): A new panel re-runs your model once for each change actually available to you — spending less, claiming later, retiring later, saving more — and ranks them by how much each moves your projection, in years of funding or dollars at 100. Levers you cannot pull are left out rather than shown greyed: no claiming row once a benefit is in payment, no retire-later row once you have retired. These are modeled outcomes under your own assumptions, not recommendations, and the sliders above the panel still let you try any size of change rather than the fixed step each row uses.
  • Fixed (Return And Volatility Now Describe One Portfolio): Expected return and return volatility were separate inputs that knew nothing about each other, so it was possible to model a 4% return carrying 12% volatility — a bond portfolio with equity-sized swings — and the simulation would answer precisely for a portfolio that does not exist. Volatility now follows the growth rate until you set it yourself, and if the pair looks unusual the calculator says so without blocking you. Nothing changes for the 8% / 12% default this ships with.

v2.2.0-beta1 — August 2026

  • Added (Why The Model Runs To 100): The model has always projected through age 100, and never said why. Seeing a plan marked as running dry at 96 reads like a failure when you don't know that 100 is a deliberately hard test rather than a life expectancy. Your results page, the assumptions table, the Methodology page and your PDF report now all explain it in plain language, using your own odds of reaching 95 rather than a generic figure — planning through 100 is about making sure a long life is something the projection absorbs, not something it depends on you not having.
  • Added (Longevity Figures In Your PDF Report): The odds of you — or, for a couple, at least one of you — still being here at 85, 90, 95 and 100 were only ever shown on screen. Your PDF report now carries them too, set directly against how many of the 300 sample runs still had money at each of those ages. Two different questions that share an age, answered side by side: how likely you are to be there, and how likely the money is.

v2.1.6-beta1 — August 2026

  • Improved (Your PDF Report Now Says What It Leaves Out): The printed report is the thing you hand to a spouse or an accountant, and it has to stand on its own — there's no page to click through to once it's on paper. Its assumptions table used to say "Medicare costs included: Yes", which is true of what we charge and misleading about what you'd actually pay. It now reads "Part B premium and IRMAA surcharges only", and the report carries a short note naming what isn't in there: the Part D plan premium, Medigap, and Medicare Advantage, with a reminder to add your own figure to your monthly spending.
  • Why this one is on the report itself: everything else the model simplifies is listed on the Methodology page, and the report points you there rather than repeating it — that way the two can't drift apart. We made an exception here because this omission is a recurring monthly cost sitting right beside a number it changes the meaning of. Nothing about your projection changes; only what the report tells you about it.

v2.1.5-beta1 — August 2026

  • Disclosed (What Medicare Costs We Don't Include After 65): Once you're on Medicare the model charges the Part B premium and both income-related surcharges — but not the Part D premium itself, and not Medigap or a Medicare Advantage plan. That's worth knowing precisely, because it isn't small: a Part D plan is commonly around $40 a month and Medigap is often more than the Part B premium. Add your own figure for these to your monthly spending, because the model won't add it for you.
  • Why we're not just estimating it: what you'd actually pay depends on the plan you choose and where you live, and the national average is a figure nobody is actually charged. Every other Medicare number in the model is the real published one, and we'd rather keep it that way and tell you what's missing than quietly blend in an average. This is now stated on the Methodology page, both in the Medicare section and in the list of what the model doesn't cover.

v2.1.4-beta1 — August 2026

  • Disclosed (Municipal Bond Interest And Your Social Security): If you hold municipal bonds, the interest is free of federal income tax — but under IRS rules it still counts towards the income figure that decides how much of your Social Security is taxable. Buying munis doesn't shelter your Social Security. Our model leaves income you've marked "not taxable" out of that figure, so if a meaningful part of your income is municipal bond interest, the tax we show on your Social Security reads low.
  • Why we're telling you rather than fixing it: the "not taxable" option also covers Roth withdrawals and the return-of-capital part of an annuity, and those correctly stay out of that figure. Splitting the option apart would put a question in front of everyone to serve the few people it applies to, and it's a question most would answer wrong. So it's now stated on the Methodology page and beside the field itself. If this affects you and you'd rather we modeled it properly, tell us — that's the signal we'd act on.

v2.1.3-beta1 — August 2026

  • Fixed (Ohio Plans Were Modeled With Too Little State Tax): This matters if you ran a plan before this release and live in Ohio. Ohio charges a flat $332 once your taxable income passes $26,050, on top of the 2.75% rate above that figure. The model was applying the 2.75% but not the $332, so Ohio households had their state tax understated by that amount in every year of the projection. Re-run your plan for a corrected projection.
  • Please note this correction goes the other way: the two fixes in the releases below found the model being harder on you than the rules require. This one is the opposite — an Ohio plan has been reading slightly better than it should. On our Ohio test household the correction added about $6,500 to the modeled lifetime state tax and brought the projected depletion age forward by around five months. We would rather tell you that plainly than quietly leave a number that flatters the answer.

v2.1.2-beta1 — August 2026

  • Fixed (Married Couples Were Taxed At Single-Filer State Rates): This matters if you ran a plan before this release, told us you were planning as a couple, and live in California, New York, New Jersey or Oregon. Those four states give married couples filing jointly wider tax brackets than someone filing alone, and the model was applying the single-filer brackets to everybody. Couples in those states had their state tax overstated for every year of the projection. In our California test household the modeled lifetime state tax fell by about $93,000 and the projected ending balance rose by about $112,000; in the New York household, about $14,000 and $17,000. Re-run your plan for a corrected projection.
  • Note on the other states we support: we checked all thirteen against their own published rate schedules, not against assumptions. Virginia, Ohio and South Carolina each publish a single rate schedule that applies to everyone regardless of filing status, so nothing changes for them — that isn't an oversight, it's what those states actually do. Massachusetts, Michigan, North Carolina, Georgia, Arizona and Colorado charge one flat rate, so filing status can't affect the result. And the joint brackets aren't simply double the single ones everywhere: California and Oregon double exactly, New York doesn't, and New Jersey has a tax band for couples that has no equivalent for single filers at all.

v2.1.1-beta1 — August 2026

  • Fixed (Married Couples Were Taxed On Too Much Of Their Social Security): This matters if you ran a plan before this release and told us you were planning as a couple. How much of a Social Security benefit counts as taxable income depends on your total income measured against a threshold, and that threshold is higher for a married couple filing jointly than for someone filing on their own — $32,000 and $44,000 rather than $25,000 and $34,000. The model was using the single-filer figures for everybody. Couples therefore had more of their benefit treated as taxable than the rules require, paid more modeled federal tax, and finished with a lower projected balance. Re-run your plan for a corrected projection; if you were planning on your own, nothing about your results changes.
  • What this changes on your results: the correction runs in your favour, and the size of it depends on how much of your income is Social Security. Across our internal test households the modeled lifetime federal tax fell by between $429 and $21,900, and projected ending balances rose by between $2,300 and $33,500. Households whose income is mostly Social Security see the largest change. Modeled depletion ages, where a plan showed one, may move later.
  • How this was found: we've begun checking the model's tax and benefit formulas directly against the published government rules they implement, rather than only against our own expectations. This was the first thing that check turned up. The arithmetic around it was correct — the two income bands, and the rule that no more than 85% of a benefit is ever taxable, all behaved properly. Only the threshold for married couples was wrong, and that is the kind of error no amount of internal testing can find, because the model was faithfully computing the wrong rule. The base amounts are set by statute and have not changed since 1993. Source: IRS Publication 915 and 26 U.S.C. § 86(c). Kessavo is not affiliated with or endorsed by any government agency.

v2.1.0-beta1 — August 2026

  • New (Already Receiving Social Security): A beta tester told us their household is already drawing benefits, in their eighties — and found the calculator had no way to say so. There's now a tick box on the Social Security step, for you and for a spouse separately. Tick it and the claiming-age question disappears, because it isn't a decision you still have, and you enter one figure instead: the monthly amount arriving now. The model pays that amount as you entered it.
  • Fixed (Benefits Already In Payment Were Being Reduced Twice): This matters if you ran a plan before this release while already receiving Social Security. The amount field asked for your benefit at full retirement age, and the model then applied the standard early-claiming reduction for whatever claiming age was selected. A benefit already in payment has that reduction in it, so entering the amount you actually receive had it taken off a second time — around 7% too low at claiming age 66, and up to 30% too low at 62, for every year of the projection. Your Social Security income, ending balance and depletion age would all have read low. Re-run your plan with the new tick box for a corrected projection.
  • Changed (Claiming Comparisons Where There Is No Choice Left): With a benefit already in payment, the claiming-age comparison — the nine lines on the balance chart, the highest-modeled-balance figure, the claiming-age slider — has nothing to compare. Those surfaces are now hidden rather than showing nine identical projections, and a shortfall is reported through the spending lever alone instead of suggesting a later claiming age that isn't available to you. The PDF report follows the same rule.
  • Note on spousal amounts: a benefit in payment already includes any spousal top-up that was applied when it was awarded, so the model no longer derives a second one for that person. What we do not model for a benefit in payment: your primary insurance amount is never asked for, so the survivor floor for a benefit claimed early cannot be applied, and full retirement age is modeled at 67 rather than the 65-to-66 range that applies to people born before 1955.

v2.0.0-beta1 — August 2026

  • New (The Years Before You Retire): If you're still working, the calculator now models what you keep saving between today and retirement. Enter your salary and the share of it you put away, and contributions compound into your accounts month by month, respecting the 2026 IRS limits — including the larger catch-up amounts from age 50 and again from 60 through 63 — and stopping in your retirement year. Previously those years did nothing at all: balances grew on investment return alone, which valued years of saving at zero for anyone not yet retired. If you've already retired, leave the salary blank and nothing changes.
  • Note for anyone who ran a plan before this release: if you were still working, your projection will now look considerably better than it did. That isn't a change of assumptions in your favour — it's the model no longer ignoring the money you're still putting away.
  • New (Social Security Estimated From Your Salary): Beta testers told us plainly that nobody casually knows what their benefit will be decades out. If you enter a salary and haven't filled the benefit in yourself, the calculator now estimates it using the published bend-point formula and fills it in for you. It's clearly labelled as our estimate rather than a Social Security Administration figure, and the moment you type your own number from ssa.gov it takes over permanently.
  • New (Other Income and One-Time Amounts): Add part-time or consulting work, rental or business income, and annuities, each with its own start age, end age, inflation treatment and tax treatment. Separately, add one-time amounts — an inheritance, downsizing proceeds, a business sale, severance, shares vesting — with the year they arrive and the account they land in.
  • New (Save and Resume): The wizard now remembers what you've entered, so a refresh or a break part-way through no longer means starting over. Everything stays in your own browser on your own device and is still never sent to us; there's a "Clear saved entries" link below the calculator whenever you want it gone. The privacy policy has been updated to describe this precisely.
  • New (Guidance on the Spending Number): Testers told us the monthly living-expense figure — the number the whole answer hangs on — was being asked for with no explanation. It now says what belongs in it, and, more importantly, what to leave out: health cover, long-term care, debt payments and income tax are all added on top by the model, so including them would have counted them twice. There's also a starting point for anyone who hasn't worked the number out yet.
  • New (More Levers on Your Results): The "Adjust your plan" controls now include your retirement year and, if you've added part-time work, how much of it — so a plan that comes up short is something you can work on from the results page rather than only read about.
  • Fixed (Spouse Claim-Age Tile): The "Highest modeled claim age — spouse" tile was appearing for everyone, including people planning on their own, for whom there is no spouse figure to compute. It now appears only when you've told us you're planning as a couple, and the spouse claiming-age slider beside it works for every couple.
  • Verification note: the Social Security bend-point formula, the early-claiming reduction and the delayed retirement credit were all checked against a Congressional Research Service worked example and reproduce it exactly, including the statutory rounding. Contribution limits come from IRS Notice 2025-67. Kessavo is not affiliated with or endorsed by any government agency.

v1.2.0-beta1 — August 2026

  • Changed (Whole Calculator Available To Everyone): Every input in the wizard is now available to everyone, with no upgrade prompts along the way — couple mode with spousal Social Security, survivor modeling, capital-gains cost basis, Roth conversion modeling, and long-term-care and debt inputs.
  • Fixed (Understated Accuracy Claims): Three notes told readers their results were less precise than they actually were. State income tax has always been modeled with each supported state’s real progressive brackets, not a flat rate; the Monte Carlo simulation has always run the full 300 scenarios, not a single path; and longevity has always been drawn from the SSA Period Life Table, not an estimated stand-in. The copy now describes what the model actually does.

v1.1.1-beta1 — August 2026

  • Improved (Monte Carlo Range Testing): The PDF report's Monte Carlo section now leads with your projected solvency range in plain language, with explanatory text under each chart; the raw trial-count tally moved to a "Professional Reference" block at the end of the section, framed for a CFP or other advisor reviewing the report with you. The online dashboard's Monte Carlo explanation was also expanded to clarify what "Success rate" means.

v1.1.0-beta1 — August 2026

  • New (Monte Carlo Range Testing): Monte Carlo results (Worst 10% / Median / Best 10%) now show the modeled funded-through age instead of a $0 ending balance when a scenario depletes before 100 — plus a funded-probability-by-age milestone breakdown and a balance-range trajectory chart, on both the online results page and the PDF report.

v1.0.1-beta1 — August 2026

  • Fixed (Couple Mode IRMAA): Medicare IRMAA surcharges for married couples are now checked against the married-filing-jointly MAGI thresholds (roughly double the single-filer thresholds, per 20 CFR § 418.1120) instead of the single-filer thresholds applied to household income. This previously overstated Medicare costs for couples with combined income above roughly $109,000.

v1.0.0-beta1 — August 2026

  • New (Public Beta Opens): Kessavo opens for public beta — a single-scenario retirement and Social Security projection through age 100 with real SECURE 2.0 birth-year RMD logic (73/75), 2026 IRS federal tax brackets, and stacked tax/IRMAA charts.
  • New (Deeper Household Modeling): Spousal Social Security (couple mode), 300-run Monte Carlo simulation ranges, SSA life-table longevity probabilities, peak tax year and IRMAA cliff detection, capital-gains cost-basis tracking, and exportable PDF plan reports.
  • New (Methodology): Methodology page — transparent formulas and citable IRS/SSA/CMS source references published alongside the calculator.
  • New (Data Privacy): All calculation inputs and scenarios run 100% locally in your browser.

What's next