Verifiable Agent Identity
Every AI agent has a cryptographically verifiable identity — who it is, on whose behalf it acts, and what authority it holds. Defined by DIDs, self-certified claims, and registry-backed verification.
The Standard for
Interoperable Intelligence
AIMCIS defines the meta-cognitive interoperability layer for global AI systems — the rules, identity, provenance, and enforcement mechanisms that make intelligence trustworthy across every domain.
About the AIMCIS Foundation
Building the standards for interoperable intelligence.
As AI systems begin to collaborate across devices, platforms, organizations, and nations, humanity needs a shared foundation — a way to ensure that every action, every transformation, and every decision made by an AI agent is verifiable, accountable, and interoperable.
The AIMCIS Foundation exists to define those rules. We develop open, enforceable standards for AI identity, provenance, policy, and secure collaboration. We publish domain-specific profiles for financial, medical, governmental, enterprise, personal, and public-information contexts. And we provide neutral governance to prevent vendor capture and ensure long-term stability.
This is the foundation for the next era of interoperable intelligence.
Four Core Primitives
The foundational guarantees that every AIMCIS-compliant system must provide.
Cryptographic Provenance & Lineage
Every decision, transformation, and state change is signed, traceable, and auditable. Ed25519 signatures form an unbroken chain from origin to current state — tamper-evident by design.
Policy & Consent Enforcement
Machine-readable rules that cannot be bypassed by vendors, operators, or agents. Consent propagates across agent hops — redaction, permissions, and usage rights enforced at every link.
Secure Session & Key Models
Ephemeral, least-privilege controls for sensitive operations. Hardware-backed key storage (HSM), session-scoped permissions, and automatic key rotation ensure enterprise-grade security.
The Twelve Domains
A complete framework covering every dimension of trustworthy AI operation.
IDNT
Identity & Entity Verification
How intelligent systems identify themselves, authenticate actions, and prove authority.
PRVN
Provenance & Lineage Integrity
Cryptographic recording of every action, transformation, and state change.
CNSN
Consent, Privacy & Personal Data Governance
Permissions, privacy, redaction, and user-centric data rights across all AI interactions.
TRNS
Transactional Integrity & Value Exchange
Guarantees for financial transactions, commitments, asset transfers, and non-repudiation.
MDCL
Medical, Clinical & Health Safety
Clinical decision support, health data protection, and medical-grade assurance.
GVRN
Governmental & Regulatory Compliance
Public records, legal processes, and civic systems requiring high accountability.
BNSP
Business & Enterprise Operations
AI within enterprise workflows, contracts, supply chains, and organizational intelligence.
INFO
Public Information & Knowledge Integrity
Authenticity, traceability, and integrity of public information and media content.
SFTY
Safety, Security & Risk Mitigation
Safety constraints, threat models, red-flag behaviors, and enforcement boundaries.
CRTV
Creative, Cultural & Recreational Systems
Creative works, licensing, entertainment, and cultural interoperability.
ENVR
Environmental & Infrastructure Systems
IoT, robotics, energy systems, transportation, and physical-world AI interactions.
AUXL
Auxiliary & Emerging Domains
Expansion domain for emerging fields and future technologies not yet formalized.
Minimum Viable Specification
The AIMCIS Minimum Viable Specification (MVS) is the published baseline that defines what every compliant system must implement to be recognized under the standard. It covers the four core primitives, the aimcis.json schema, DNS TXT verification, HTTP header conventions, and the initial domain profiles.
What the MVS Defines
- aimcis.json — Schema v1.0 for publisher self-declaration at
.well-known/aimcis.json - DNS TXT Records — Verification format:
aimcis=pub-XXXXXXXX - HTTP Headers —
X-AIMCIS-Publisher,X-AIMCIS-Certification - Core Primitives — Agent Identity, Provenance Chain, Policy Rules, Session Keys
- Domain Profiles — All twelve domains with AL0–AL3 assurance criteria
- Certification Framework — Badge system, AAA attestation, revocation rules
aimcis.json Schema (v1.0)
{
"$schema": "https://aimcis.org/schema/aimcis.json/v1.0",
"publisher_id": "pub-XXXXXXXX",
"domains": ["IDNT", "PRVN", "TRNS"],
"certification_level": "AL0",
"registered_at": "2026-10-26T00:00:00Z",
"dns_txt_verified": true,
"principles": {
"agent_identity": true,
"cryptographic_provenance": true,
"policy_enforcement": true,
"secure_sessions": true
}
}
Assurance Levels
| Level | Name | Verification | Issuer |
|---|---|---|---|
| AL0 | Aware | Self-declaration | Entity + Registry |
| AL1 | Compliant | Automated checks | Registry |
| AL2 | Verified | Third-party attestation | AAA |
| AL3 | Certified | Annual recertification | AAA (annual) |
The MVS is published and available for review.
View Full SpecificationsAdoption Lanes
Six pathways through which the ecosystem grows — each with defined milestones and sprint cadences.
Lane A
Developer Ecosystem
SDKs (Python, JavaScript, Go), CMS plugins, grant program, and annual developer conference. 90-day sprints aligned with quarterly SDK releases.
Lane B
Platform Integration
AI platform AL1+ commitments, signal parsing integration, and the "AIMCIS Platform Alliance." 60-day sprints matching rapid platform release cycles.
Lane C
Regulatory Pull
EU AI Act, W3C, and ISO PAS submissions. Regulatory recognition with enforcement references to AIMCIS certification levels.
Lane D
Policy & Government
Government agencies certified at AL2+ for GVRN, CNSN, and INFO domains. 120-day sprints aligned with procurement cycles.
Lane E
Publisher Networks
Journalism associations, CMS foundations, and academic consortia. Peer referral effects driving network-wide adoption.
Lane F
Browser Infrastructure
Native Trust Indicator integration in browsers with 1B+ combined users. Badge rendering at AL1–AL3 across all assurance levels.
Begin Your AIMCIS Journey
Whether you're a publisher, platform, browser, or AI agent — start at AL0 with self-declaration and advance through verified certification.