---

How Accurate Is OpenFate? One Birth, Five Systems, Checked Against Open-Source Software

We ran one sample birth through OpenFate's free Western, Vedic, Bazi, Zi Wei Dou Shu and Human Design charts, then recomputed every result with open-source astronomy and calendar libraries. Here is what matched, the one thing that differed, and what an audit like this can't tell you.

Independent review by Chen Tian Ji, Changsha, China. Chen is an astrology and Bazi enthusiast, unaffiliated with and unpaid by OpenFate. He checked one sample birth against the libraries listed at the end so that anyone can reproduce the results.

Tested in September 2026 on the English site, free tools only, not signed in, no credits used. The birth data is invented and belongs to no real person.

Why audit the math first

An AI astrology reading is only as good as the chart underneath it. If the Ascendant lands in the wrong sign, or the Bazi hour pillar falls in the wrong two-hour block, the explanation that follows will still be fluent and confident. It will just be about somebody else.

OpenFate is built to keep those two layers apart: system-specific engines calculate the chart, and the AI is instructed to explain those facts rather than generate them (methodology). That separation makes the calculation layer checkable from the outside, so that is what we checked.

The test

Sample birth: May 2, 1992, 4:40 p.m., Chicago, Illinois. Sex, which only Bazi and Zi Wei use (to set the direction of luck cycles), was entered as female. All settings were left at their defaults.

Two things shaped the choice. Chicago was on Central Daylight Time that afternoon (UTC−5), so a tool that ignored daylight saving would shift every time-sensitive result by an hour. And before settling on this date, we screened candidate births so that every value we planned to compare sat well away from a boundary: sign cusps, nakshatra and pada edges, Human Design line edges, the Chinese double-hour and solar-term changes. No pass or fail in this audit comes down to rounding.

The reference software

Reference	Used for	Independent of OpenFate engine?
PyEphem 4.2.1	Western planets, Ascendant, Midheaven	Yes. Different planetary theories from Swiss Ephemeris.
pyswisseph (Swiss Ephemeris)	Lahiri sidereal positions, Human Design activations	Same ephemeris family; checks implementation, not the ephemeris itself.
lunar-javascript 1.7.7	Bazi four pillars	Yes
iztro 2.6.1	Zi Wei palaces and stars	Yes
Own scripts	Dasha, Human Design channel and type, luck-pillar onset	Written from published rules.

Scorecard

System	Compared	Result
Western	Sun, Moon, eight planets, Ascendant, Midheaven	All 12 within 1 arcminute of PyEphem
Vedic	Nine grahas, lagna, nakshatras, padas, current dasha	Match
Bazi	True solar time, four pillars, luck-pillar direction and start	Match
Zi Wei Dou Shu	Life and Body Palace, bureau, 14 major stars, Four Transformations	Match except one documented school difference
Human Design	26 activations, Type, Authority, Profile, Definition	Match

Not checked: the Sukuyo birth mansion (no independent reference we trusted), the oracle tools, and anything written by AI.

Western astrology: twelve positions, all within an arcminute

The free Western chart shows UTC-5 at birth and a tropical zodiac.

Body	OpenFate	PyEphem
Sun	Taurus 12°43′	12°43.1′
Moon	Taurus 14°46′	14°46.0′
Mercury	Aries 17°30′	17°30.5′
Venus	Taurus 1°32′	1°32.2′
Mars	Pisces 27°41′	27°41.9′
Ascendant	Libra 6°36′	6°36.4′
Jupiter	Virgo 4°38′	4°38.3′
Saturn	Aquarius 17°57′	17°57.2′
Uranus	Capricorn 17°57′	17°57.9′
Neptune	Capricorn 18°54′	18°54.9′
Pluto	Scorpio 21°49′	21°49.5′
Midheaven	Cancer 7°43′	7°44.0′

OpenFate shows whole arcminutes and appears to truncate rather than round, which is why Mars at 27°41.9′ displays as 27°41′. The Midheaven's one-minute gap sits within the precision of the simplified obliquity formula in our own check; Swiss Ephemeris gives 7°43′.

The result page also has a collapsed panel, “Chart settings and calculation checks,” that states the convention behind every number: tropical zodiac, Placidus houses, mean lunar node, Swiss Ephemeris, and a line reporting that planetary positions were cross-checked “within the recorded tolerance against Astronomy Engine.” Chart settings offer seven house systems, true or mean nodes, adjustable aspect orbs and four optional asteroids.

Vedic: sidereal positions, nakshatras and the dasha clock

The Vedic chart declares a Lahiri sidereal zodiac with whole-sign houses. On this date the Lahiri ayanamsa is about 23.75°, so each sidereal position should equal the tropical one minus that amount. All nine grahas matched in sign, degree (to the 0.1° shown), nakshatra and pada. The lagna is Virgo 12.9° in Hasta; the Moon is Aries 21.0° in Bharani, pada 3.

The dasha timeline is a good test because it compounds several steps. A Moon in Bharani starts life in a Venus mahadasha; with the Moon 7.7° into that 13°20′ nakshatra, about 42% of Venus's 20 years remained at birth. Running the Vimshottari sequence forward from there puts this chart in a Rahu mahadasha from October 2023 to October 2041, with a Rahu–Jupiter antardasha from July 2026 to December 2028. OpenFate shows exactly those dates.

One more built-in check: in Ashtakavarga, the Sarvashtakavarga bindus always total 337. OpenFate's table sums to 337.

Bazi: solar time and the four pillars

Bazi depends on local solar time rather than clock time. Chicago lies about 2.4° east of the 90°W meridian that Central time is based on, which adds roughly 9.5 minutes; the equation of time on May 2 adds about 3.3 more. Take off the daylight-saving hour, and 4:40 p.m. CDT becomes about 3:53 p.m. local solar time. OpenFate shows “True Solar Time 15:52,” within a minute of our figure and well inside the 3– 5 p.m. Shen hour.

The free Bazi chart shows true solar time and the four pillars.

The pillars, Ren-Shen, Jia-Chen, Wu-Yin and Geng-Shen, are identical to lunar-javascript's. The luck pillars run backward, as they should for a woman born in a yang year. OpenFate dates their start to October 17, 2001. Counting back to the Qingming solar term (12:45 UT on April 4, 1992) and applying the classical rule that three days equal one year gives the same date. The decades line up as well: Xin-Chou for 2021–2030 and Geng-Zi for 2031–2040.

Zi Wei Dou Shu: the one difference, and why it isn't an error

The Zi Wei chart places the Life Palace at Wu-Shen with no major star in it, the Body Palace at Ren-Zi in the Career Palace, and the chart in the Earth Five bureau. iztro agrees on all three, and all 14 major stars sit in the same palaces in both charts. OpenFate labels stars in English, so the Emperor Star is Zi Wei and the Treasury Star is Tian Fu.

Branch	Palace	Major stars
Shen	Life	none
Wei	Siblings	Zi Wei, Po Jun
Wu	Marriage	Tian Ji
Si	Children	none
Chen	Wealth	Tai Yang
Mao	Health	Wu Qu, Qi Sha
Yin	Travel	Tian Tong, Tian Liang
Chou	Friends	Tian Xiang
Zi	Career (Body)	Ju Men
Hai	Property	Lian Zhen, Tan Lang
Xu	Spirit	Tai Yin
You	Parents	Tian Fu

The Four Transformations for a Ren year are where the two charts part ways. Both put Lu on Tian Liang, Quan on Zi Wei and Ji on Wu Qu. For Ke, OpenFate marks Tian Fu, while iztro's default marks Zuo Fu. This is a long-running disagreement between schools. Wang Tingzhi's Zhongzhou school assigns the Ren-year Ke to Tian Fu, and his own writing explains why; OpenFate declares Zhongzhou as its standard. The difference is a stated convention rather than a miscalculation.

Human Design: all 26 activations

A Human Design chart has two sets of 13 activations: Personality, at the birth moment, and Design, at the moment the Sun was 88° of solar arc earlier (here, about 07:09 UT on February 4, 1992). We recomputed all 26 and mapped them onto the 64 gates and six lines. Every one matched OpenFate as gate.line, from the Personality Sun at 24.6 to the Design Sun at 13.2.

From there the derived keys follow. Only one channel is complete, 19–49, joining the Root and Solar Plexus. The Sacral is undefined and no motor connects to the Throat, which makes the chart a Projector. The defined Solar Plexus gives Emotional authority; the Personality and Design Sun lines give a 6/2 profile; the two defined centres form a single definition. OpenFate reports all four, and it shows the evidence next to each one rather than just the label.

One detail matters if you compare apps. The Human Design activations matched only when we used the true lunar node. With the mean node, the node gates move (10.6 instead of 10.4 on the Personality side, 58.5 instead of 38.1 on the Design side). OpenFate's Western chart defaults to the mean node, and so does its Vedic Rahu. Each choice is common in its own tradition, but node conventions are a real source of cross-app differences.

What this audit shows, and what it doesn't

It shows: For this birth, OpenFate handled daylight saving, located the city, converted to local solar time where needed, and produced chart facts in five systems that match third-party references to the precision displayed. The one difference traced back to a declared school convention.

It does not show: One sample cannot prove every birth, date range, or edge case is handled correctly. It does not show that astrology, Bazi, Zi Wei, or Human Design predict anything. It does not test AI readings, which use credits.

Other things worth knowing before you try it

Strengths. Each result names the conventions behind it: zodiac or ayanamsa, house system, time zone, solar-time correction, school. Chart calculation is free in all six systems without an account. Notation is explained where it trips people up; the Human Design page notes that “24.6 means Gate 24, Line 6—not a score.”

Limitations. The pages are dense, and the free charts already carry plain-language guidance beside the data. It is labeled, but readers who only want the raw chart will scroll past a lot of it. Positions are truncated to whole minutes. Node conventions differ between systems, as noted above. The quick form asks for an exact birth time; the chart pages describe how unknown times are handled (the Vedic page, for example, says it suppresses the lagna, houses and dashas), but we didn't test that path.

Cost and data. According to OpenFate's FAQ, full AI reports, compatibility readings and typed follow-up questions use credits, and the cost is shown before you confirm. A 250-credit pack is currently US$6.99, and one-time credits don't expire. Subscriptions renew automatically, and cancelling takes effect at the end of the billing period. Eligible refund requests must be made within 7 days (see the pricing page). OpenFate states that it does not sell personal information. AI memory is on by default and can be turned off or cleared in settings (privacy policy).

Should you try it?

If you want a free chart you can inspect, in one tradition or several, the calculation layer held up under this audit, and the settings panels tell you which convention you are looking at. A sensible order is to start with the system you already know, check that chart against a tool you trust, and only then decide whether an AI explanation of it is worth paying for. That last question is one this audit doesn't answer.

Reproduce this audit

Input: May 2, 1992, 16:40 local time (21:40 UT), Chicago, Illinois (41.88°N, 87.63°W); female where asked; default settings. OpenFate pages: /en/astro-chart /en/vedic-chart /en/bazi-chart /en/ziwei-chart /en/human-design-chart References: PyEphem 4.2.1 (apparent geocentric ecliptic longitudes of date; Ascendant and Midheaven from local sidereal time); pyswisseph with Lahiri ayanamsa and true node; lunar-javascript 1.7.7 with local solar time as input; iztro 2.6.1 (astro.bySolar('1992-5-2', 8, 'female'), the Shen hour). Human Design gate order: 64 gates of 5.625° each, starting with Gate 41 at 2° Aquarius; six lines of 0.9375° per gate.

Sources and further reading

Accessed September 2026. References include OpenFate calculation validation and free charts, PyEphem, Swiss Ephemeris, lunar-javascript, iztro, and Wang Tingzhi on the Ren-year Ke transformation. All screenshots are from OpenFate. Official website: OpenFate.ai