The Pipeline: Question In, Reading Out
A traditional Qimen consultation has four stages: fix the moment, cast the chart, find the question in the palaces, interpret. A master does them with an almanac, brush and thirty years of pattern memory. QimenIt does the same four stages in software: the first three deterministically, the last one with AI. Nothing in the pipeline is a black box, and this article walks through each stage.
Step 1: Pinning Down the True Moment
Everything downstream depends on the hour, and the hour depends on the sun, not the clock. QimenIt computes your true solar time from your location: four minutes of correction per degree of longitude, with daylight saving removed where it applies. Near a two-hour boundary, or the midnight day-change, this correction alone decides which chart you get. (The full story is in True Solar Time Explained.)
Step 2: Casting the Chart
With the moment fixed, casting is pure rule-following. The solar term and day select one of the 18 classical ju: nine yang configurations for the waxing year, nine yin for the waning half. The earth plate seeds the palaces; the nine stars, eight doors and eight deities then rotate along the Luoshu path into their positions for your hour. Two practitioners (or two computers) following the same school's rules will produce identical charts. This is arithmetic, and computers simply do not make arithmetic slips at 11:40 p.m. A general chatbot re-derives these steps by prediction on every run; how often that lands, and why it still cannot be relied on, is unpacked in Can ChatGPT Read a Qimen Chart?
Step 3: Locating Your Question in the Chart
Classical practice assigns every question type a set of use-gods (用神), the symbols that stand for the matter at hand. A job offer lives with the Open Door; a payment you're chasing lives with the Life Door; a partnership lives with Partner; you yourself live with the day stem. QimenIt reads your question, classifies it, and pulls the relevant palaces, the same triage a master performs by eye, encoded as a lookup rather than a judgment call.
Step 4: Interpretation
Only now does AI enter. The model receives the computed chart, the located palaces, and your exact wording, and weighs the factors a human reader weighs: element tiers, seasonal strength, the relationships between your palace and the question's palace, and known symbol combinations. The output is a reading in plain language, with the chart attached.
The division of labor matters: AI never "decides" what chart you have; the chart is computed before the model sees anything. AI only does what humans also do with a finished chart: read it. If you took the same chart to a teacher, you could check every line.
Where AI Beats a Human Master, and Where It Doesn't
- Exact solar time: computed astronomically, never estimated from a paper table
- Zero casting errors: the mechanical steps are mechanical
- No motivated reading: it doesn't soften bad news for a paying client, or pad sessions
- Always available: 3 a.m. decisions get the same quality as 3 p.m. ones
- Consistent school: one rule set, applied identically every time
- Reading you, not just the chart: hesitation, history, what you didn't say
- The follow-up question: a live teacher probes until the real question surfaces
- Lineage nuance: rare combinations passed down inside schools, thinly documented
- Accountability: a person who remembers you and revisits the call
The practical conclusion: AI is the strongest first reader Qimen has ever had: flawless mechanics, instant availability, and no incentive to tell you what you want to hear. For most everyday decisions, that is the whole job. For the biggest calls of a life, a chart in hand makes a conversation with a human teacher better, not redundant; you arrive with the mechanics already done. How this first reader actually scores — the honest numbers and how we validate them — is laid out in How Accurate Is a Qimen Reading?
See the pipeline run on your own question. Your first reading is free, and the chart, the located palaces, and the solar-time correction are all shown with it.
Cast a Free Chart →Frequently Asked Questions
How does AI read a Qimen chart?
In four stages: convert clock time and location to true solar time, cast the chart by classical rules, locate the use-gods (用神) for the question type, then interpret. The first three stages are deterministic computation; AI performs only the final, interpretive step.
Does AI decide which chart I get?
No. The chart is computed before the model sees anything. AI only does what a human reader also does with a finished chart: read it.
Will two people asking at the same moment get the same chart?
Yes. Casting is pure rule-following: the solar term and date select one of the 18 classical ju, and the plates rotate into place. Two practitioners — or two computers — following the same school's rules produce identical charts.
Is an AI reading better than a human master?
They are complements. AI wins on precision, patience, and price: exact solar time, zero casting errors, no motivated reading, available at any hour. A human teacher wins on context, follow-up questions, lineage nuance, and accountability.