addyears.ai

1 of 17

No account needed. Everything stays on your device — no data is sent to any server. We may ask for your email later, but only for our optional longevity newsletter. Completely voluntary.

Your estimated healthspan

Every choice you make from today can move this number.

One last thing

Stay on track with a weekly nudge.

Every week — one piece of longevity science, a reminder to log, and a snapshot of how your number is moving. No noise. Unsubscribe any time.

Estimated healthy life remaining

—

YEARS

—

MONTHS

—

DAYS

This week

—

earned this week

day streak

days logged

The compounding effect

Pattern detected

Your weekly insight

🔒

Log for 3 more days to unlock your first personalised insight.

7-day healthspan gain

Today's factors

Log today to see your breakdown.

How we calculate

Addyears uses published epidemiological data — not guesswork. Here's exactly what drives your number and how confident we are in each piece.

Your starting number

Your baseline healthspan estimate is built from three inputs: country healthy life expectancy (HLE), adjusted for biological sex (~3 years on average for females, based on WHO Global Health Observatory data), and a lifestyle score derived from your onboarding answers.

We use HLE rather than total lifespan because what matters is the number of years spent in good health — mobile, cognitively intact, and independent. HLE data is sourced from WHO country profiles and national health registries (2022–2024).

The lifestyle score adjusts the HLE estimate up or down based on your reported habits, with each factor weighted according to its published hazard ratio. The result is a population-level estimate, not a personal prognosis. Think of it as the starting point for a conversation, not a verdict.

How daily logs move the number

Each daily log calculates a composite hazard ratio (HR) — your estimated relative mortality risk compared to a reference person with average habits. An HR below 1.0 means lower risk than the reference (a gain in healthy time); above 1.0 means higher risk (a cost).

We multiply the HRs for each factor you log into a single composite HR, then apply an overlap discount of 35%. This discount acknowledges that lifestyle factors are correlated — a person who exercises also tends to sleep better and eat well — so the combined effect is not simply multiplicative. The 35% figure is a conservative estimate drawn from multi-factor interaction analyses.

delta = (1 − composite HR) × remaining years × 0.35 ÷ 365

The result is a daily delta in minutes or hours. A very good day might add 15–25 minutes of expected healthy life. A poor day might cost 30–50 minutes. These are intentionally modest — that is what the evidence actually supports at the single-day level. The power comes from consistency over months and years.

Primary data sources

Exercise — UK Biobank cohort (Ding et al., 2022 · N=366,000)

Moderate MVPA vs. inactive: HR 0.88 · High MVPA: HR 0.81 · Combined cardio + strength: HR ~0.80 · Sedentary reference: HR 1.10. Published in British Journal of Sports Medicine.

Smoking — Jha et al., NEJM 2013 (N=1.48M)

Lifetime daily smoker all-cause mortality: HR 2.50 · Ex-smoker: HR 1.35. Daily log uses conservative per-day estimates (heavy day HR 1.35, light day HR 1.18) since single-day HRs cannot be directly derived from lifetime cohort data.

Sleep — Cappuccio et al., Sleep 2010 (79 cohorts, N=1.5M)

Short sleep <5h: HR 1.30 · 6h: HR 1.14 · 7h: HR 1.00 (reference) · Long sleep ≥9h: HR 1.34. Dose-response across the full range.

Diet — EPIC-elderly (Knoops et al., JAMA 2004) & Nurses' Health Study

Full Mediterranean dietary pattern HR 0.77 for high adherence vs. low. Daily log estimates are conservative (HR 0.88–0.93 per day) — the published HR requires sustained adherence over years, not a single meal.

Alcohol — GBD 2016 Alcohol Collaborators, Lancet 2018

Heavy drinking (>4 drinks/day): HR 1.80 · Moderate (1–2/day): HR ~1.04–1.08 · Rare or none: HR 0.97. Note: the "J-curve" protective effect of light drinking remains contested; we use conservative estimates.

Social isolation — Holt-Lunstad et al., PLOS Medicine 2015 (N=1.3M)

Social isolation: HR 1.33 · Loneliness: HR 1.26 · Adequate social connection: HR 0.95 (Million Veteran Program). Effect size comparable to smoking 15 cigarettes per day.

Sitting time — Biswas et al., Annals of Internal Medicine 2015 (N=595,000)

Prolonged sitting ≥10h/day: HR 1.34 · Effect is independent of leisure-time exercise — standing up matters even if you go to the gym.

Stress — Kivimäki et al. & multiple cohorts

Chronic high stress: HR 1.28 (cardiovascular mortality). Work stress meta-analysis (Kivimäki, Lancet 2012, N=200,000): HR 1.23 for CHD. We use HR 1.12–1.28 across the stress spectrum.

Cold exposure — Laukkanen et al., JAMA Internal Medicine 2015

Frequent sauna use (4–7×/week): HR 0.94 for all-cause mortality (Finnish cohort, N=2,315). Cold water immersion evidence is emerging and more limited — we apply conservative estimates (HR 0.985–0.995 per session) until larger RCTs are available.

Dog ownership — Mubanga et al., Scientific Reports 2017 & AHA Statement

Dog ownership associated with ~24% lower cardiovascular mortality (HR ~0.95, Swedish registry N=3.4M). Mechanisms: increased physical activity, social interaction, stress reduction.

Supplements — Multiple RCTs and meta-analyses

Vitamin D: VITAL trial (N=25,000) — 12% reduction in cancer mortality. Omega-3: ASCEND trial — modest cardiovascular benefit. Evidence for creatine, NMN/NAD+, magnesium in longevity is preliminary. Total supplement benefit capped at HR 0.95.

Genetics vs. lifestyle — the real picture

"Genes account for perhaps 25% of the variation in human longevity. The rest is environment and lifestyle."

Christensen et al. · Nature Reviews Genetics · 2006

The foundational evidence comes from Danish twin studies (Herskind et al., 1996, N=2,872 twin pairs), which established that genetic factors explain approximately 25% of lifespan variation, with environment and lifestyle accounting for the remaining 75%. This figure has been broadly replicated across multiple populations.

A 2023 study in Nature Medicine (Youssef et al.) analysed UK Biobank data and found that lifestyle factors collectively explained far more variance in biological ageing than genetic risk scores — even among those with high polygenic risk for age-related disease.

A 2022 study in Lancet Healthy Longevity (Li et al., N=11,941, 20-year follow-up) found that a healthy lifestyle offset the genetic risk of shorter lifespan — participants with poor genetics but good habits lived significantly longer than those with good genetics and poor habits.

What this means: your number in addyears.ai is not fixed by your genes. Epigenetic research (including the work of Steve Horvath on biological age clocks) confirms that lifestyle interventions measurably reverse biological ageing markers. The baseline can move. That is the entire point of this app.

Honest limitations

This is not medical advice. Addyears is a personal reflection tool grounded in population-level epidemiology. It cannot account for your individual genetic makeup, medical history, or unmeasured environmental factors. Always consult a doctor for health decisions.

Population HRs ≠ individual predictions. A hazard ratio of 1.28 for chronic stress means that in large cohorts, stressed individuals die about 28% sooner on average — not that you specifically will. Individual risk varies enormously around any population mean.

Single events vs. sustained patterns. One Mediterranean meal does not deliver the same benefit as years of dietary adherence. We use conservative per-event estimates in the daily log, but the published HRs for diet and exercise reflect long-term habits. The real benefit compounds with consistency.

Hazard ratios are not fully independent. Multiplying HRs assumes independence between factors — exercise, diet, sleep, and stress are correlated in real populations. Our 35% overlap discount partially corrects for this but is an approximation, not a precise statistical adjustment.

Cold exposure and some supplements have limited RCT evidence. We use the most conservative estimates available and will update values as stronger evidence emerges.

The projection is probabilistic. "Likely healthy to ~83" reflects the central tendency for people with your profile in population data — not a personal forecast. Think of it as a weather forecast for a population, not a diagnosis for an individual.

Data & privacy

All your data is stored locally on your device. Nothing is sent to any server. Addyears has no account system and cannot see your logs.

Delete all my data

Removes your profile, all logs, and resets the app completely. This cannot be undone.

What if…?

Curious what a choice adds or costs in healthy time? Pick anything below to see the science behind it.

ℹ️

This is not your log. Exploring here doesn't record anything. To track your actual day, tap the + button in the centre.