How a birth chart is calculated

Start with where the planets actually were

A birth chart begins with an astronomical question: at this instant, from this point on Earth, where were the Sun, Moon and planets? That question has one correct answer, and it does not depend on any astrological opinion.

This engine answers it from the JPL DE440s kernel — the 1849–2150 segment of NASA/JPL’s DE440 planetary ephemeris, the same solar-system model used for spacecraft navigation. Positions are computed geocentrically and verified to arcsecond precision. A birth date outside that 1849–2150 window is outside what the kernel covers.

No language model is anywhere near this step. Positions are computed, not generated.

Subtract the ayanamsa to get sidereal positions

Those positions are tropical: measured from the vernal equinox, which drifts slowly against the stars because the Earth’s axis precesses. Vedic astrology works sidereally, measuring from a fixed point against the star background instead. The difference between the two is the ayanamsa.

This engine uses the Lahiri ayanamsa, anchored at 23.856453° — 23°51′23.2″ — for J2000, with the precession in longitude added for any other date. Subtract that from a tropical longitude and you have the sidereal longitude a Vedic chart is built from.

The choice of ayanamsa is a real fork: a different one shifts every position by a fraction of a degree, which is usually invisible by sign and occasionally decisive at a boundary. Which one is in use is stated on your chart page rather than left implicit.

Divide the circle into 27 nakshatras

The 360° circle is divided into 27 equal nakshatras of 13°20′ each. A planet’s nakshatra is simply which of those 27 arcs its sidereal longitude falls in, and how far through that arc it has travelled.

That second number — the fraction of the nakshatra already elapsed — is not decoration. The Moon’s fraction is what sets the starting point of the entire Vimshottari dasha timeline, so it propagates into every date the chart later reports.

Whole-sign houses: one sign, one house

The ascendant is the sidereal longitude rising on the eastern horizon at the birth moment, which is where the birth time and the birthplace enter the calculation.

This engine uses whole-sign houses, the oldest and simplest of the house systems: the sign containing the ascendant is the first house in its entirety, the next sign is the second, and so on. In code it is one line — the house number is the distance between a planet’s sign and the ascendant’s sign, plus one.

This is worth knowing because house systems disagree. Under a quadrant system such as Placidus, a planet can sit in a different house than it does here. Neither is a computation error; they are different definitions, and this one is stated rather than assumed.

Why the same details always give the same chart

Every step above is a deterministic function of the birth moment and place. There is no sampling, no model, and no randomness anywhere in the chain, so the same birth details produce the same chart to the arcminute — today, next year, on any machine.

That is a low bar in principle and an uncommon one in practice, and it is the reason the chart and the interpretation are kept strictly apart. The chart is a computation you could in principle redo yourself. What it means is a separate question, answered from cited sources.

What the birth time actually changes

The ascendant moves roughly one degree every four minutes, so an uncertain birth time moves the houses first. A twenty-minute error can shift the ascendant by about five degrees and, near a sign boundary, into the next sign entirely — which rotates every house in the chart.

Planetary positions by sign are far more forgiving: over twenty minutes the Moon moves about a sixth of a degree and the slower planets barely at all. So an approximate birth time still gives you reliable placements by sign, and unreliable houses. The chart page shows exactly what was used, so you can change it and watch what moves.