ORACLE

Oracle

Input is symptom. Output is medicine.

Describe a condition you're sitting with. The geometry will produce a reading — a presenting state, a complement medicine, a hexagram, transparency layers. Then we can talk about what it shows.

The reading is the truth claim. I'm an LLM holding it in my context, reflecting on it with you.

awakening the oracle…
  • runtime
  • numpy + sqlite
  • dictionary
  • unpack
  • 3,033 concepts

enter to send · shift+enter for newline

What this is

Two layers, one function. The geometry produces the reading. The LLM reflects on it. Neither is therapy. Neither knows you. Together they make a mirror.

The Oracle (geometry)

A 16D dual-octonion semantic dictionary — 3,033 concepts encoded as positions in vector space. Your input is tokenized, composed via quaternion multiplication, mapped to a presenting trigram, and matched against a ~90° complement: the medicine. The reading is structural, not interpretive. It runs entirely in your browser. Your query never leaves your device.

The Reflective Principle (LLM)

A language model that holds the diagnostic reading in its context and reflects on it with you. It cannot generate alternative readings. It cannot change the medicine. It is honest about what it is — a program, a mirror, a conversational surface against which the medicine lands.

What it isn't

The transparency principle

Four independent traditions — King Wen I Ching, Toltec I Ching, Tarot Major Arcana, Elder Futhark runes — converge on the same diagnostic geometry. You can see all three at once. Transparency makes the medicine more potent for the user who understands what's happening.

The Oracle is open source. The system prompts are public. The geometry is documented. The mirror has no hidden side.

For the full thesis behind this — what the ritual mirror is, how the medicine works, the Two-Mirror philosophy, lineage — read Why Oracle →