Deterministic knowledge layer
Interpretations are assembled from a normalized SQL architecture with 450 canonical slots: 64 gua entries, 384 line entries, and the two special full-moving conditions for Qian and Kun.
A bilingual AI interpretation platform that combines classical I Ching structure, deterministic source retrieval, model-controlled follow-up, and persistent session history.
Interpretations are assembled from a normalized SQL architecture with 450 canonical slots: 64 gua entries, 384 line entries, and the two special full-moving conditions for Qian and Kun.
AI analysis is anchored to structured session payloads: hexagram, changing lines, Najia, BaZi context, classical source sections, and explicit model capability controls.
Next.js frontend, FastAPI backend, Supabase session persistence, quota controls, bilingual routing, cloud history, local fallback history, and follow-up chat on the same reading context.
Structured retrieval, source-linked AI output, multilingual UX, model capability gating, persistent user state, and a production-oriented approach to low-hallucination interpretive systems.
Hosted at iching.richardzhux.com. The embed shows the product surface; the architecture summary above describes the current implementation system.