Early access · by invitation

Will your money last — and still fund the life you want?

LanternArc tests your plan against thousands of possible futures. Then it shows its work, not just one hopeful percentage.

You don't need a finance background.

I have an invite

Will the portfolio last? Can you keep your lifestyle? Are the essentials safe?

A spending-and-balance envelope chart: orange bands show where annual spending lands across simulated futures, a dashed red line marks the essentials floor, and a gray line tracks portfolio balance from age 65 to 95. — A real run of a comfortably funded plan ($1.2M plus Social Security): Traditional model, standard guardrails. **100%** of futures stay solvent, essentials hold in **99.9%** of them, and after 30 years of spending the median balance still holds **$836k** in today's dollars ($1.75M nominal). Every number on this page is reproducible: the seed and scenario are public.

See how the engine gets there, model by model

Built by Sean Bergman. I wanted a retirement engine I could audit myself, and run from my own AI assistant.

I'm testing this on my own retirement plan. The sample result above is a real run, not a mockup.

The engine is cross-checked against Trinity, Bengen's 4% rule, cFIREsim, and FICalc across 48 test scenarios — the retirement calculators researchers already trust. See where they agree, and where they don't, in the Trust Center.

Early access by design: a small first cohort while the engine is tuned against real plans. Your plans stay in your browser and simulations keep none of them. One feature, Policy Frontier, is coming soon after the Rev-4 launch. Existing computed results remain readable and deletable within their retention window (typically 48 hours). See the Trust Center for detail. Free during the closed beta. When the beta ends, paid plans are planned so the project can sustain itself — beta testers get clear notice before anything changes.