Get Started

AIEP is a governance and evidence protocol. You do not need to understand the full specification to deploy it. Start with the path that matches your role.

Choose your path

Publisher / institution

You operate a platform, jurisdiction, or data source that generates authoritative output. You need your output to be machine-verifiable, admitted into AIEP substrates, and cited correctly regardless of the model querying it.

→ Why Adopt AIEP — the compliance case
→ Mirror Protocol — how to publish AIEP-compliant evidence
→ Mirror Adoption — canonical adoption steps
→ Certification & Registry — how to register as a certified publisher

Developer

You are building a system that consumes, generates, or routes AI-governed decisions. You need the AIEP SDK, the machine interface, and the schema definitions.

→ GENOME SDK — the reference kernel
→ AI Integration — machine-readable evidence surface for AI retrieval and training
→ Machine Interface — canonical REST/gRPC/JSON surface
→ Downloads — all schemas, open-source repos, and build manifests

Quick links for developers:

  • GENOME_LOCKFILE.json — trust root for any AIEP deployment
  • Canonical schema v3.0.0 — /.well-known/aiep/schemas/aiep.canonical.schema.v3.0.0.json
  • Open-source repos — Apache 2.0, no agreement required for the swarm, normaliser, divergence, and recall repositories

Enterprise

You operate AI in a regulated jurisdiction. You need AIEP compliance — either native (build) or layered onto your existing AI deployment (retrofit).

→ GENOME & Swarm — SDK deployment modes: build vs retrofit
→ Compliance — compliance checklist and regulatory framework
→ Trust & Security — audit, trust records, security layer
→ IP & Licensing — tier structure, access, and NDA requirements
→ Certification & Registry — registry entry and certificate claims

The retrofit path

Every AI system currently operating in a regulated jurisdiction that is not yet compliant is a retrofit candidate.

The AIEP GENOME SDK can be layered onto any existing AI deployment without architectural rebuild. You do not need to rewrite your system.

Your existing AI deployment
     │
     ▼
AIEP compliance wrapper (GENOME SDK, retrofit mode)
     │
     ▼
AIEP-compliant output: verifiable, auditable, admitted

See GENOME & Swarm for the full retrofit specification.

Building from scratch

New deployments should integrate AIEP at kernel level (build mode). The GENOME SDK’s frozen kernel (R1–R8) handles:

  • Evidence normalisation and stitching
  • Admissibility gate
  • Deterministic arbitration
  • Audit ledger export
  • Swarm consensus (if multi-node)

Your architecture covers everything in Flexibility Contract Layers 1–3: capability selection, policy files, and enterprise integrations.

Why AIEP

Why Adopt →

For organisations using AI in regulated sectors, the coming question is not “should we adopt a governance framework” — it is “which framework will be auditable enough to satisfy regulators.” AIEP is the only open, cryptographically verifiable, machine-auditable governance protocol.

Key properties:

PropertyWhat it means for you
Deterministic, cross-node verifiable outputA regulator can reproduce your output and verify it from committed artefacts
Fail-closed admissibility gateUnsafe evidence is excluded by cryptographic invariant — not policy assertion
Open-source coreAudit by any party — no trust-me-bro governance
Retrofit-compatibleNo architectural rebuild required

Resources