← Free Kundli · Accuracy check · September 2026

Our maths was checked against NASA's ephemeris

Every astrology site says its calculations are “accurate”. None of them says how accurate, or compared to what. This page gives you the numbers — whatever they turned out to be.

What was compared with what

Our site calculates in your browser, using VSOP87 for the planets and ELP2000 for the Moon. These are accepted astronomical theories.

The same instants were then calculated separately using NASA JPL's DE440s ephemeris — the same file used to navigate spacecraft — together with NAIF SPICE kernels and IERS Earth-orientation data, in a completely separate Python program.

Two independent routes, one question: where was the planet at that moment?

Result — 1 September 2026, 12:00 IST, Delhi

What that actually means

One nakshatra pada is 3°20′ — that is 12,000 arcseconds. The largest difference we found was 12.1 arcseconds.

For comparison: if your recorded birth time is off by just one minute, the chart moves about 75 times more than that. Which means changing the ephemeris at this level would gain nothing. The room for improvement is in how precisely your birth time is known — not in the maths.

Time handling — 12 hard cases

Planetary positions alone are not enough. The real errors usually happen when converting local time to UTC. So 12 deliberately awkward cases were built, with their correct answers computed in advance from DE440s. Our engine got 12 out of 12 right, including:

• Kolkata, 1900 — before India had standard time at all; correct offset +05:21:10
• Delhi, 1943 — the wartime +06:30
• Kathmandu — the quarter-hour +05:45
• Kiritimati — UTC+14, where the date itself runs ahead
• And three local times that never existed — the engine now refuses them instead of quietly substituting some other moment.

Several large public astrology APIs accept only a numeric UTC offset. Kolkata 1900 or Delhi 1943 simply cannot be expressed in such an API — those cases don't fit its model.

Every kundli comes with its own working

Below every result on Free Kundli there is a small panel — “How this was calculated”. It lists the things that are usually nowhere to be found: which ayanamsa and its exact value, which house system, the exact UTC instant derived from your local time, and the source of the planetary maths.

The benefit is direct: when two websites disagree, you can see whether the difference comes from a convention or a mistake. Without these details that difference cannot be understood at all.

And now the part that keeps this page honest

Everything above says only one thing: that the positions were computed correctly. It does not say that any life outcome attributed to those positions is scientifically established.

A planet's position is measured. Its meaning comes from tradition. These are two different things, and conflating them is the most common mistake in this field — showing precision and passing prediction off as science.

That is why we publish the arcsecond figures but sell no guarantee alongside them.

Know the limits too

• The DE440s kernel only covers 1849 to 2150. For an earlier date we have nothing to compare against (and the site itself does not generate charts outside 1900–2100).
• The ayanamsa is mean Lahiri — 23.8531° at J2000, carried forward and back by the Meeus precession polynomial. It is a convention, not a scientifically settled number.
• Houses are whole sign only (rashi = house). Placidus, Bhava Chalit and others are not implemented.
• This comparison was made at one instant, not at every instant. Differences will vary slightly on other dates — but will still be tiny next to a pada.

Next: Create your free kundli ·