Concept 概念
How Pillarwise works: every line traced to its source
Most AI astrology hands you a paragraph and asks for trust. Pillarwise hands you the trail instead — the chart fact, the rule it fired, and the classical line it was transcribed from.
Pillarwise works by tracing every line of a reading through a fixed chain: a fact in your chart triggers a named rule, and that rule is transcribed from a public-domain classical text. A deterministic engine computes the chart; governed rules map its patterns to plain-English tendencies; the wording only phrases what the engine already derived. Nothing appears that cannot be traced back to a chart fact and a cited source.
Reasoning first, wording second
A Pillarwise reading is the output of a fixed process, not a paragraph a model improvised. A deterministic engine computes your Four Pillars, derives the Day Master, Ten Gods and strength relationships, and applies governed rules that map those patterns to plain-English tendencies. Only then is the result put into words. Because the engine runs first, the same birth details always yield the same chart and the same set of tendencies — there is nothing to drift or embellish.
The chain behind every line
Each line of a reading carries its own trail. Expand it and you see three linked things: the chart fact that set it off, the rule that fact triggered, and the classical source that rule was transcribed from.
| Layer | What it is | Example |
|---|---|---|
| Chart fact | A verifiable feature of your computed chart | Day Master is Yin Metal, weak for the season |
| Rule (rule_id) | A governed mapping from that fact to a tendency | dm.metal.yin.weak → resourceful-but-guarded tendency |
| Classical source | The public-domain text the rule was transcribed from | 三命通會, quoted in the original beside the English |
Where a classical line reads starkly, it is kept as the language of its era and framed plainly — a tendency described in old idiom, shown beside the modern reading so you can weigh the source yourself.
Why this is different from black-box AI astrology
| Typical black-box AI astrology | Pillarwise | |
|---|---|---|
| Where the words come from | A model generates fluent text | A fixed engine derives findings; the model only phrases them |
| Consistency | Same details can produce different readings | Same details always produce the same reading |
| Sources | Unnamed, generic archetypes | Named public-domain classical texts, cited per line |
| What you can check | The paragraph, take it or leave it | The chart fact, the rule, and the quoted source |
| Register | Confident claims about outcomes | Tendencies to reflect on — never a verdict |
Why traceability is the point
Traceability is what lets you disagree well. When you can see that a line rests on a specific pattern and a specific text, you can decide how much weight it deserves in your own situation. A reading you can interrogate is a reading you can actually use — you are reflecting on evidence, not accepting an assertion. This is decision support for self-knowledge, offered against your season, your context, and your own judgment.
Questions 問答
Frequently asked
Yes. Every interpretation expands into its chain: the chart fact that triggered it, the rule that fired, and the public-domain classical text the rule was transcribed from, quoted in the original alongside the English.
No. A deterministic engine derives the chart and its tendencies first. Language models only phrase findings the engine already produced — they do not invent interpretation of their own, and the same birth details always produce the same reading.
Public-domain sources such as 三命通會, 淵海子平 and 子平真詮. Each rule is cross-checked against the wording it was transcribed from, so the reasoning is checkable rather than asserted.
No. Every line is framed as a tendency to weigh against your season, context, and judgment — decision support for self-reflection, not a statement about the future.
Related 相關