Three systems, strictly divided
Meridian practices Chinese metaphysics — not one tool repeated three times. Zi Wei Dou Shu maps life domains and role structure, and how the organizing center of the current decade is shifting. Qi Zheng Si Yu reads the long-term operating mechanisms shown by the birth sky — how real-world position forms, step by step, through information, value, resource, and contract. Qi Men Dun Jia reads the entry point, the gaps, and the sequence of action at the moment of the question. The three answer different layers of the question, not the same question three times.
Each brief uses only the tools it actually applied, and states what each contributed, where they converge, and where they diverge. All are traditional interpretive frameworks, not predictive science: they give you more structures to reason with — not a fixed future.
The chart calculation is deterministic
The charts themselves (Zi Wei palaces and transformations, Qi Zheng planetary positions, the Qi Men configuration) are computed by fixed rules — not guessed by a language model. The calculation method, time-zone and true-solar-time handling, treatment of an uncertain birth or question time, and the version are recorded and can be checked.
Corrected birth data invalidates the outputs that depend on it and produces a new version with the date and reason.
Historical calibration is a fit gate
Before a full reading, the chart is calibrated against things that actually happened to you. This is a test of data quality and case fit — not scientific validation, and not proof of accuracy.
Calibration records matches, misses, and ambiguity. If it doesn’t hold, the case does not proceed under the published acceptance rule, and the deposit is returned.
Software assists, and it is disclosed
Software may assist with deterministic calculation, organization, translation, formatting, consistency checks, and scheduling. It is never described as an autonomous master, and its drafts are never delivered without human review.
Every delivery states which tasks were software-assisted. Your data is not used to train any model by default.
Human review and the case record
Chart inputs, interpretation, the written brief, and substantive revisions must be reviewed by a person. Each case carries a case number, reviewer-role ID (e.g. GF-R01), method version, review date, and revision history.
Meridian does not publish a personal name or photograph, and does not invent a master persona. Accountability belongs to this traceable process and record.
The exact schools and calculation rules are still being finalized and will be published here as the version updates.