Transparency
Method, sources and where AI stops
Agua separates four layers that are easy to blur together: the classical text, our translation, a modern explanation, and question-specific advice. Each is labelled on the page.
The chart is computed by rules
Casting, line values, the resulting hexagram, the stem-branch assignment, the six relations, the self/other pair, strength ratings and timing are all produced by deterministic code. The same values and timestamp always give the same chart.
The AI layer never casts, never changes a line value, and never alters the rule conclusion. It rewrites the wording of the seed advice so it fits your question, and answers follow-ups about the same chart.
Texts and translation
The classical judgments, images and line texts are public-domain Chinese sources and are shown as they are, in Chinese, beside our translation.
Translations and explanations on this site are our own, produced with AI assistance and reviewed in-house. They are not the work of an external academic translator, and modern published translations are not copied.
Where AI stops, and what you can verify
AI output is validated before it is shown: it must match the chart's facts, keep the same lean as the rule result, and avoid absolute promises. When validation fails, the page keeps the rule-based reading and says so.
If the AI service is unavailable, every traditional function still works. Nothing about the chart depends on a model being reachable.