narrative-intelligence: A three-universe processing framework that analyzes events through Engineer, Ceremony, and Story Engine perspectives, performs narrative coherence analysis, classifies emotional beats, and provides unified narrative narrative state management with NCP schema and Redis-backed persistence.
View narrative-intelligence on npmjs.com
The ava-langgraph-narrative-intelligence package enables developers to create narrative-aware processing systems that analyze events through three interpretive universes:
NarrativeCoherenceEngine scores narrative health across structural, thematic, character, sensory, and continuity dimensions, identifies gaps with severity and routing targets, and produces a Trinity assessment (Mia/Miette/Ava8 synthesis)EmotionalBeatClassifierNode classifies story beats across 10 emotional tones (Devastating, Hopeful, Tense, Joyful, Melancholic, Triumphant, Fearful, Peaceful, Conflicted, Resigned) with keyword scoring and confidence metricsUnifiedNarrativeState) across six systems in the stack, with Redis-backed persistence through NarrativeRedisManager and episode lifecycle managementThe current reality provides complete three-universe processing, coherence analysis, emotional classification, and state management. The creative tension lives in advancing toward structural thinking — a four-node graph (picture→draft→review→revise) that ports miaco’s pde-to-st command into a LangGraph subgraph — and toward consent-gated episode retrieval that respects relational accountability when surfacing historical narrative context.