KPR Simulator

The problem

A KPR, an Indonesian home mortgage, is the largest financial commitment most households will ever make. Banks quote a fixed rate for two or three years, and then the loan floats. A family that budgets around Rp3.2 juta a month can face Rp4.8 juta in year four. Every calculator I found assumed one rate for the whole twenty years, so it hid the biggest risk in the loan.

The calculators had other gaps too. "Flat" and "efektif" interest are both advertised, they can't be compared directly, and most borrowers can't tell which one they are being quoted. The calculators show a number with no working. And nearly every one in Indonesian search results exists to collect leads for a bank or a property portal.

What I built

The page answers the question a borrower can take to the bank

I changed what the calculator asks. "Can I afford Rp3.2 juta?" becomes "My loan breaks at, say, 13.4%. My bank doesn't publish its margin over the base rate, so what is it?" The threshold page finds the exact floating rate at which the instalment passes a share of income you choose. Two edge cases get their own plain answers: already over the limit at a zero rate, and still affordable above the highest rate searched.

Every rupiah can be traced

Money is always a whole number of rupiah. Rounding is its own labelled step in the working, and it names the convention it follows. That step is where calculators drift a few rupiah from bank statements, so I made it visible. The engine never reads today's date; the month is always passed in. The same inputs give the same result forever, which matters when someone checks it against a bank statement six months later.

Two separate calculations check each other

The effective rate is worked out twice, by two routines that share no code. If they shared a helper, they could share a bug. This check also corrected my own brief. The PRD said a flat 5% is roughly a 9% effective rate. That holds over one to five years, but over twenty years the computed figure is 7.95%. The app states the computed number, and I fixed the brief.

Unknown figures are refused by name

Every regulated figure, such as the subsidy rate, price ceilings, income ceilings and tax rates, is stored with its legal basis, source link and verification date. The build fails if one has no citation. When I couldn't trace loan-to-value ratios, the minimum down payment or the subsidised loan tenor to a regulation still in force, I didn't guess. They became seven declared gaps the app names and explains in both languages. Amber marks the floating-rate band and every assumed value, and red marks the affordability limit. Nothing else uses those colours.

Your figures never leave the browser

Income and loan amounts are stored in the part of the link after the hash sign, which browsers never send to a server. The site makes no network requests and has no analytics or ads. A build check confirms the exported site calls nothing external.

Result

KPR Simulator is live, public and free of ads, in Indonesian first and English second. It shows the full payment schedule with an amber band where the rate becomes unknown. It covers the threshold rate, flat and efektif side by side, and the subsidised FLPP path with a check that names which criterion fails. It adds total cost of ownership and an extra-payment simulator. A parameter page lists all 27 cited figures. I built it alone over two days in August 2026.

The app never advises, ranks or names a bank. It states in the interface that it is not financial advice, that the floating rate is unknown and unpublished, that bank fees vary, that loan approval is not modelled, and that figures should be confirmed with the bank.