Do Leap Months Break Your Saju Chart?
No — and the proof is a leap month that changes its own month pillar halfway through.
Bokki Editorial
The short answer
No. Saju's engine never checks whether a birthdate fell inside a lunar leap month, because it never reads the lunar calendar's month number at all. Every pillar Bokki computes — year, month, day, hour — comes from getFourPillars(), which starts from a solar (Gregorian) date and looks up where that date sits against the 24 solar terms (節氣). The lunar calendar's leap-month flag, the thing that makes 윤달 a leap month in the first place, is a completely different piece of information that this pipeline never asks for. So a birth on a leap-month date computes exactly the same way as a birth on any other date: no leap-specific check, no leap-specific branch, nothing.
The worry that's actually real — an ambiguous lunar record
The genuine problem isn't in the math — it's in old records. If all you know is "born on the 15th of the 2nd lunar month, 2023," that sentence has two different real answers, because 2023 had both a regular 2nd month and a leap 2nd month (윤2월). The regular 2nd month's 15th day fell on the solar date 2023-03-06; the leap month's 15th day fell on 2023-04-05 — a full month apart. Run both through Bokki's engine and you get genuinely different charts: 2023-03-06 gives a month pillar of 甲寅, while 2023-04-05 gives 乙卯. If a birth record only says "음력 2월15일" without marking which one, there is no way to compute the right answer from the sentence alone — you would need to know which of the two real days actually happened. That's a record-keeping ambiguity, not a flaw in how Saju is calculated.
Proof: one leap month, two month pillars
Here is the sharpest way to see that Bokki's calculation is blind to leap-month status specifically, not just to the lunar calendar in general. The 2023 leap 2nd month (윤2월) runs, without a single missed day, from the solar date 2023-03-22 through 2023-04-19 — by the lunar calendar's own reckoning, that whole 29-day stretch is one unbroken month. But the solar term Qingming (清明) falls on 2023-04-05 at 09:13, right in the middle of it, between the leap month's 15th and 16th days. Bokki's month pillar follows Qingming, not the lunar month boundary: day 15 (2023-04-05) still reads 乙卯, and day 16 (2023-04-06) already reads 丙辰. One single lunar month — leap or not makes no difference here — ends up carrying two different Saju month pillars, because the calculation was never tracking the lunar month to begin with. It was only ever tracking solar terms.
Why the ambiguity in §2 never reaches Bokki
Bokki's forms never ask for a lunar birthdate anywhere — check saju, compatibility, daily readings, day-picking, all of it: every birthdate field on this site is solar (Gregorian) only, and there is no lunar-to-solar conversion screen. That means the exact ambiguity described above — "2월15일, but which one?" — never has a chance to enter the pipeline, because Bokki was never asked to resolve it. Once you know your solar birthday, whatever lunar month or leap status that day happens to correspond to is simply irrelevant information the calculation was never going to use. The source code says this plainly: the function that reads a birth month's pillar is commented as computing "the same thing getFourPillars's birth month pillar uses — the solar-term month changes not on the 1st, but the instant a 節 arrives," with no mention of the lunar calendar at all.
What we don't calculate
We don't offer a lunar birthdate field, so if the only record you have is an ambiguous lunar date like the one in §2, Bokki cannot resolve which of the two solar days is the real one — no calendar library can recover information a record never wrote down, and we say so rather than guessing. We don't implement the separate folk custom of treating leap months specially for choosing auspicious dates (건축·이장 등을 윤달에 하는 관습) — that belongs to day-picking traditions, not to Saju's pillar arithmetic, and Bokki's own day-picking tool doesn't reference leap-month status either. We don't adjust traditional Korean age-counting for leap years. And nothing here ranks a leap-month birth as luckier or less lucky than any other (D-003) — a month pillar is arithmetic against a solar-term calendar, and arithmetic doesn't carry an opinion.
Closing note — one leap month, two pillars, zero leap-specific code
In short: a leap-month birth computes exactly like any other birth, because Saju's month pillar comes from crossing a solar term (節氣), and the lunar calendar's leap-month flag is never read anywhere in that path. The clearest proof is the 2023 leap 2nd month itself — one unbroken 29-day lunar month, 22 March through 19 April — which still splits into two different Saju month pillars, 乙卯 and 丙辰, the moment Qingming falls between its 15th and 16th days. The real difficulty people run into is a record problem, not a math problem: "음력 2월15일, 2023" names two different solar dates a full month apart (03-06 versus 04-05, 甲寅 versus 乙卯) depending on whether it means the regular month or the leap month, and no calculation can recover which one without the birth record actually saying so. Bokki sidesteps the whole question by never asking for a lunar date in the first place.