What does the math, and where.
Every calculator is deterministic Python in backend/calculators/. No LLM. The same inputs always produce the same outputs — which means findings are reproducible and disputable, not black-box.
FIRE
Five lifestyles. Five targets. Five dates.
Lean / Base / Comfort / Coast / Loan-free. Each runs a year-by-year simulation with step-up SIPs, inflation-adjusted (real) returns, LTCG drag with the ₹1.25L exemption, and the loan-free variant uses your weighted real loan rate instead of a heuristic.
Tax
Old vs new regime, every year.
FY-aware. Computes both regimes side-by-side, applies LTCG with the ₹1.25L exemption (Budget 2024), classifies senior citizen status as of 31 March of the relevant FY, and surfaces the regime delta in rupees, not percentages.
Net Worth
Joint loans deduplicated. Property apportioned correctly.
Joint home loans appearing in two spouses’ bureaus are deduplicated by account number in the family view. Property equity is computed as ownership% × (value − linked_loan × ownership%) so co-owners share both the asset and the debt.
HLV
Human Life Value, with a working rule-of-thumb formula.
Sajag uses a working heuristic: max(income × 10, income × years_to_retire × 0.6) + outstanding loans. Actuaries use several methods (income replacement, expense replacement, capital-needs); the heuristic is what almost every Indian advisor anchors on. Compared to your current term cover. Health cover compared against a 10× monthly-expense baseline. Gaps surface as findings; ULIPs / endowments get an IRR verdict (surrender / paid-up / continue).
Money Health
A 100-point composite across 10 pillars.
Savings rate, emergency fund, term cover gap, health cover gap, debt service ratio, retirement-on-track, asset diversification, productivity of property, credit utilisation, and tax efficiency. Each pillar is 10 points; the score is brutally honest, not encouraging.
Five scenarios. Real returns. No magic.
The FIRE engine runs five lifestyle variants against your household’s actual numbers. For each, the target corpus is inflated_annual_expense / SWR, default safe-withdrawal-rate 3.5% — conservative because Indian post-tax bond yields are lower than the textbook US assumption.
Corpus growth is a year-by-year simulation, not a closed-form shortcut. Step-up SIPs apply at year boundaries. Returns use the Fisher equation: (1 + nominal) / (1 + inflation) − 1. LTCG drag is applied only to real gains above ₹1,25,000 a year. The Loan-free FIRE variant uses your balance-weighted loan rate from the bureau report, not a textbook 9% guess.
The brutal-honesty engine.
Money leaks are the things every household tells itself a small lie about. The engine looks for specific shapes: a credit card with a revolving balance, a ULIP where the IRR-after-charges beats nothing, a Coast-FIRE timeline that doesn’t pencil out, a tax regime choice that costs the household ₹40k a year.
Every finding has a quantified rupee impact and a one-line recommendation. Nothing is “might want to consider.” The point is to make the lie visible and the fix obvious.
Sample finding
Credit card revolving — paying 36%+ interest
Critical · ₹17,630/yrPay the full ₹48,972 balance from next paycheck. Never revolve a CC; the interest more than wipes out any reward points.
The promise, in numbers.
10 minutes to your FIRE date: enter age, target retirement age, monthly expenses, monthly savings. The calculator answers in seconds. Refining the answer with real CAS / EPFO / NPS data takes a few uploads but doesn’t change the order of magnitude.
30 minutes to find your money leaks: upload a CAS, a bureau report, and 12 months of bank statements. The brutal-honesty engine runs in under a second once the data is in. The bottleneck is downloading three PDFs from three portals, not the math.