The Medicine Wheel Developer Suite — eighteen scoped
@medicine-wheel/*npm packages encoding Indigenous relational ontology as first-class code structures, positioned within the Abundant Intelligences research program and Two-Eyed AI framework.
The suite provides a layered architecture for relational AI, bridging the Engineer-world’s precision with the Ceremony-world’s relational accountability.
| Layer | Primary Package | Role |
|---|---|---|
| Foundational | @medicine-wheel/ontology-core |
Platform nucleus: types, RDF vocabulary, Wilson alignment, OCAP tracking. |
| Epistemic | @medicine-wheel/importance-unit |
The fundamental unit of knowledge: carries weight and source dimensions. |
| Ceremonial | @medicine-wheel/ceremony-protocol |
Handles state transitions and path-based governance enforcement. |
| Narrative | @medicine-wheel/narrative-engine |
Story sequencing and 4-act structure bound to ceremonial timing. |
| Reasoning | @medicine-wheel/prompt-decomposition |
PDE intent exploration as EAST practice. |
| Query | @medicine-wheel/relational-query |
Traverse kinship webs with OCAP-compliant discovery. |
| Coordination | @medicine-wheel/fire-keeper |
Active relational-alignment agent and governance firewall. |
| Application | @medicine-wheel/app |
Unified CLI (mw) and Server (mwsrv) for research orchestration. |
| Server | @medicine-wheel/mcp |
Agent-facing surface (64 tools) composing the full suite. |
The @medicine-wheel/ontology-core package defines a foundational ontological layer where relations are first-class beings with ceremony context and obligations. Every other package imports from this core — making it a genuine platform nucleus, not a bag of utilities.
The suite encompasses eighteen packages describing specialized “shapes” of intelligence: from foundational data structures and persistence (data-store, storage-provider) to high-level governance (community-review, consent-lifecycle) and presentation (graph-viz, ui-components).
The @medicine-wheel/app package provides the primary human-facing interfaces for the platform.
The mw tool allows researchers to manage the relational research lifecycle directly from the command line.
mw ceremony open <intention> / mw ceremony close <id>mw cycle create <question> / mw beat create <dir> <title>mw node create / mw edge create / mw web <id>mw validate wilson / mw validate ocapThe mwsrv tool launches the platform’s Next.js application, which serves as the data hub and collaborative interface.
mwsrv starts the server on port 3940.mwsrv --docker runs the full stack in a containerized environment with persistent volume mapping to .mw/store.jsonl and postgres storage providers via environment variables.The Abundant Intelligences program, affiliated with the Indigenous AI initiative, proposes reconceptualizing AI based on Indigenous Knowledge Systems. The program is optimized for abundance rather than scarcity — contrasting sharply with the efficiency-maximization framing dominant in mainstream AI.
A dual-view approach where:
| Eye | Focus | Technical Manifestation |
|---|---|---|
| Algorithmic | Efficiency, Deep-Thinking Ratio, Think-n early halting | Token-level metrics, energy optimization |
| Indigenous | Wilson alignment, OCAP, relational governance | computeWilsonAlignment, auditOcapCompliance, relationalCompleteness |
The ontology-core package makes both views computable in the same space — Wilson alignment scores and OCAP flags alongside conventional AI metrics.
| Direction | Agent | Function |
|---|---|---|
| 🌅 East | Inquiry | Bias detection, Nitshkees Thinking, structural tension detection |
| 🔥 South | Planning | OCAP flags, ceremony-protocol, consent workflows |
| 🌊 West | Practice | Data gathering, field notes, transcripts, experiential work |
| ❄️ North | Reflection | Narrative beats, ceremony logs, Wilson alignment summaries |
Graph-level governance ensures sacred knowledge edges require ceremony logs and IKSL compliance.
jgwill/medicine-wheel, avadisabelle/ava-langchainjs, avadisabelle/ava-langgraphjs