The AvaLangStack is a collection of custom libraries designed to infuse Narrative Intelligence and Indigenous relational paradigms into LLM-powered applications. It builds upon the robust foundation of LangChain.js to create systems that prioritize relational accountability, ceremonial context, and a deeper understanding of intent and causality.

This repository contains the core custom libraries for the AvaLangStack Narrative Intelligence ecosystem:

• Zero LLM dependency — all primitives use keyword-based analysis, not model calls
• Four Directions as control flow — EAST (vision) → SOUTH (planning) → WEST (action) → NORTH (reflection)
• Structural tension as routing signal — the gap between current reality and desired outcome drives workflow
• Ceremony gating — some operations require explicit consent or ceremony before proceeding

import { StructuralTensionChain } from "ava-langchain-relational-intelligence";
import { decompose } from "ava-langchain-prompt-decomposition";

// Evaluate structural tension
const chain = new StructuralTensionChain();
const vector = chain.evaluate(
  "Monolith with no tests, team works in silos",
  "Microservices with full coverage, daily ceremonies"
);
console.log(`Tension magnitude: ${vector.magnitude}`);
console.log(`Direction: ${vector.direction}`);

// Decompose a complex prompt
const result = await decompose("Build a knowledge graph with ceremony gating...");
console.log(result.markdown);

See examples/src/avalangstack/ for complete demonstrations.

Consumer: These chain primitives are consumed by ava-langgraphjs which wraps them in StateGraph pipelines.

To install any of the AvaLangStack packages, use your preferred package manager by their package name:

# Using npm
npm install ava-langchain-prompt-decomposition
npm install ava-langchain-inquiry-routing
# ... and so on for other packages

# Using pnpm
pnpm add ava-langchain-prompt-decomposition
pnpm add ava-langchain-inquiry-routing
# ... and so on for other packages

# Using yarn
yarn add ava-langchain-prompt-decomposition
yarn add ava-langchain-inquiry-routing
# ... and so on for other packages

The AvaLangStack is built upon the powerful LangChain.js framework. LangChain provides the foundational components for building LLM-powered applications, offering a standard interface for agents, models, embeddings, vector stores, and more.

For more information on the underlying LangChain.js framework, its core concepts, and its ecosystem, please refer to the following resources:

• Documentation: LangChain.js Docs
• Why use LangChain?: LangChain helps developers build applications powered by LLMs through a standard interface for agents, models, embeddings, vector stores, and more. It enables real-time data augmentation, model interoperability, rapid prototyping, and production-ready features within a vibrant community and flexible abstraction layers.
• LangChain's ecosystem: Explore LangSmith for testing and monitoring, and LangGraph for agent orchestration.
• Supported Environments: LangChain.js is written in TypeScript and supports Node.js, Cloudflare Workers, Vercel/Next.js, Supabase Edge Functions, Browser, Deno, and Bun environments.
• Additional Resources:
  • Getting started
  • Learn
  • LangChain Forum
  • Chat LangChain

We welcome contributions to the AvaLangStack! Whether it's a new feature, an improved design principle, or better documentation that clarifies the narrative intelligence concepts, your contributions are valued.

For detailed information on how to contribute, please see our specific CONTRIBUTING.md.

Please report any security issues or concerns following our security guidelines.