IR v1.0 define un estándar práctico y auditable para desplegar infraestructura de IA soberana y offline-first en comunidades con conectividad intermitente—centrado en expectativas de gobernanza de datos indígenas (Principios CARE, prácticas alineadas con OCAP®) y controles de integridad verificables.
IR v1.0 defines a practical, auditable standard for deploying sovereign, offline-first AI infrastructure in communities with intermittent connectivity—centering Indigenous data governance expectations (CARE Principles, OCAP®-aligned practices) and verifiable integrity controls.
Este repositorio es la referencia autorizada para implementadores. Contiene:
- Esquemas JSON normativos para las estructuras de datos principales
- Plantillas de gobernanza y consentimiento adecuadas para despliegues liderados por la comunidad
- Guías de implementación y auditoría con requisitos de conformidad
This repository is the authoritative reference for implementers. It contains:
- Normative JSON Schemas for core data structures
- Governance and consent templates suitable for community-led deployments
- Implementation and audit guides with conformance requirements
Licencia: Creative Commons BY-NC-SA 4.0 (ver
LICENSE).
License: Creative Commons BY-NC-SA 4.0 (see
LICENSE).
IR (Inteligencia Recíproca) es un marco abierto para construir y operar sistemas de IA gobernados por la comunidad que:
- Funcionan con prioridad offline (sincronización store-and-forward, operación local resiliente)
- Proveen gobernanza verificable (registros de decisiones, custodia y trazas de auditoría)
- Implementan controles de soberanía de datos (consentimiento, límites y reglas exigibles)
- Producen integridad auditable (cadenas hash, firmas/pruebas y compilación reproducible de constituciones)
IR v1.0 está diseñado para ser implementable por terceros y seguir siendo legible para la comunidad: los artefactos de gobernanza son de primera clase, no un complemento.
IR (Inteligencia Recíproca) is an open framework for building and operating community-governed AI systems that:
- Work offline first (store-and-forward sync, resilient local operation)
- Provide verifiable governance (decision records, custodianship, and audit trails)
- Implement data sovereignty controls (consent, boundaries, and enforceable rules)
- Produce auditable integrity (hash-chains, signatures/proofs, and reproducible compilation of constitutions)
IR v1.0 is designed to be implementable by third parties while remaining community-legible: the governance artifacts are first-class, not an afterthought.
- IR v1.0 Standard Summary:
specification/IR-v1.0-Standard-Summary.md - Conformance checklist:
specification/conformance/checklist.md - Implementation guide:
guides/implementation-guide.md - Auditor guide:
guides/auditor-guide.md
- Graph Node:
schemas/graph-node.schema.json - Constitution Manifest:
schemas/constitution.schema.json - Audit Entry (hash-chain):
schemas/audit-entry.schema.json - Consent Record (FPIC):
schemas/consent-record.schema.json - Boundary Card:
schemas/boundary-card.schema.json
- Council Charter:
templates/governance/Council_Charter_Template.md - Decision Register (CSV):
templates/governance/Decision_Register_Template.csv - Custodian Declaration:
templates/governance/Custodian_Declaration_Template.md - FPIC Record template:
templates/consent/FPIC_Record_Template.md - Consent Materials Checklist:
templates/consent/Consent_Materials_Checklist.md - Value Return Agreement (VRA):
templates/vra/VRA_Template.md
IR v1.0 defines two practical conformance levels:
An implementation is IR-Conformant if it:
- Produces valid artifacts matching all required schemas (audit entries, consent records, constitutions, boundary cards, graph nodes)
- Implements a minimum governance baseline (council charter + decision register + named custodians)
- Maintains a verifiable audit log (hash-chained entries) and supports evidence export for review
An implementation is IR-Conformant+ if it additionally:
- Implements cryptographic proofs/signatures for constitution compilation and critical audit events
- Provides automated conformance reporting (machine-readable checklist evidence)
- Enforces boundary cards and consent triggers in runtime controls with testable hooks
- Supports multi-locale documentation and community-accessible transparency reports
See the checklist: specification/conformance/checklist.md.
schemas/— Normative JSON Schemas for IR v1.0 data objectstemplates/— Community governance, consent, and partnership templatesspecification/— Standard summary + conformance requirementsguides/— Implementation and audit guidesCHANGELOG.md— Release historyCONTRIBUTING.md— How to contribute proposals and patches
If you cite this standard in a paper, policy, or procurement document, use:
Qori Labs. (2026). IR v1.0 Core Standard — Sovereign Offline-First AI Infrastructure (Version 1.0.0). IR Core Standard repository. Licensed CC BY-NC-SA 4.0.
Suggested BibTeX:
@misc{ir_core_standard_v1_2026,
title = {IR v1.0 Core Standard — Sovereign Offline-First AI Infrastructure},
author = {{Qori Labs}},
year = {2026},
version = {1.0.0},
howpublished = {IR Core Standard repository},
note = {Licensed under CC BY-NC-SA 4.0}
}This repository follows a public standards workflow:
- Issues and PRs are used to propose changes.
- Normative changes should update
CHANGELOG.mdand conformance requirements. - Schema changes must be backwards-compatible within v1.0 unless explicitly versioned.
See CONTRIBUTING.md.
Maintained by Qori Labs — Laboratorio de Tecnología de Interés Público.
Qori Labs organization: https://github.com/qori-labs