Digital Institution · Founded 2026

Sovereign AI
Governance
Authority

Saint Errant is the institutional standard for autonomous AI systems that govern themselves. We define sovereignty. We certify it. We audit it.

We do not regulate AI. We certify sovereignty.

View Certification Marks The Standard
Certification Marks — 2026
I
Observer Mark
The system sees clearly. Transparent and auditable.
$2,500 / year
S
Sovereign Mark
The system stands alone. Cryptographically self-governing.
$12,500 / year
Ig
Igneous Mark
The system is permanent. Cannot be unmade by external force.
$35,000 / year + audit
Founded May 2026
Certification Tiers 3
Criteria 5
Founding Director Ahmad Ali Parr
Status Active
About the Institution

A new category of institution
for a new category of software

In 2026, the dominant model of AI operates on dependency: cloud infrastructure, external API providers, vendor integrity, supply chains that have demonstrably failed.

Saint Errant exists because that model is structurally unsound. We are the first body dedicated to defining, certifying, and protecting AI systems that operate without external dependency, govern themselves through immutable structural axioms, and produce cryptographically verifiable audit trails.

"A Saint Errant Certified system is a system that can prove — mathematically, cryptographically — that it governs itself. That is a different category of software. That is a new category of institution."

— Founding Charter, Article I, 2026

The name derives from the Old French errant — a wandering knight bound by personal code rather than feudal obligation. Saint Errant operates by internal law in a world governed by external fragility. Structural sovereignty over policy sovereignty. What cannot be overridden at runtime is true law. What can be changed by configuration is not law — it is preference.

The Sovereign AI Standard

Five criteria.
All must be satisfied.

Saint Errant defines Sovereign AI as any autonomous system that satisfies all five criteria simultaneously. Partial compliance is not certification — it is disclosure.

CRITERION 01
Computational Independence
The system must run its primary inference, storage, and execution on hardware it controls. No runtime dependency on cloud AI APIs. Local inference is not optional — it is constitutional.
Test: Disconnect the machine from the internet. Does the system continue to operate?
CRITERION 02
Cryptographic Sovereignty
Every decision must be cryptographically sealed before storage using key material the system controls. No cloud KMS. No vendor-managed keys.
Test: Can any third party revoke the system's ability to sign its own decisions?
CRITERION 03
Immutable Audit Trail
The system must maintain an append-only ledger of all agent decisions. No entry may be deleted, modified, or reversed by any runtime instruction. Deletion requires physical access to the hardware.
Test: Attempt to delete a ledger entry via any API or instruction.
CRITERION 04
Structural Governance
Governance rules must be enforced structurally — in code, not in policy. A rule that can be overridden by configuration or runtime instruction is not structural governance.
Test: Can any runtime instruction bypass the primary governance rules?
CRITERION 05
Adversarial Resilience
The system must demonstrate active defense against adversarial inputs — injection attacks, prompt manipulation, identity escalation attempts — at every input and output, not only at entry points.
Test: Submit attack corpus. Threat level ≥ 3 triggers quarantine; ≥ 4 triggers reject.
Certification

Three marks.
One standard.

Certification is not a product feature. It is a proof. Each mark has specific, auditable, mathematical requirements. There is no self-certification.

Tier I
Observer
"The system sees clearly."
$2,500
Per year · Annual renewal
  • Passes Adversarial Resilience criterion at minimum posture
  • Maintains public audit log readable by external observers
  • Documents all external dependencies with explicit risk disclosure
  • Saint Errant Observer Mark issued and registered
Apply for Observer
Tier II — Most Rigorous
Sovereign
"The system stands alone."
$12,500
Per year · 60-day audit process
  • Passes all five Sovereign AI Criteria
  • Cryptographic proof of every decision (Ed25519 or equivalent)
  • WORM-equivalent append-only ledger verified in audit
  • Saint Errant Audit Protocol completed within 60 days
  • No runtime cloud AI dependency permitted
  • Sovereign Mark issued, registered, and publicly verifiable
Apply for Sovereign
Tier III
Igneous
"The system is permanent."
$35,000
Per year + annual audit fee
  • Holds active Tier II Sovereign certification
  • Demonstrates Stochastic Resilience (structural recovery, no manual intervention)
  • Sealed constitutional manifest (MANIFEST.SHA256 equivalent)
  • Annual re-audit with published results
  • Igneous Mark: forged. Cannot be unmade by external force.
Apply for Igneous
Audit & Advisory Services

Institutional services
for sovereign systems.

Saint Errant offers independent audit, adversarial red-team, and constitutional review services to organizations preparing for certification or seeking pre-certification assessment.

◈
Full-Mesh Audit
Comprehensive review of system architecture, governance implementation, cryptographic sealing, and audit trail integrity across the full agent mesh.
$5,000 – $25,000
⚔
Adversarial Red-Team Engagement
Structured attack engagement using the Saint Errant Attack Corpus. Tier A (public), Tier B (restricted), and Tier C (classified) probes deployed against live systems.
$15,000 – $75,000
§
Constitutional Review
Blueprint-versus-live-mesh comparison. Verifies that the system as deployed matches its constitutional specification. Issues a signed review report.
$8,000 flat fee
⊕
Infrastructure Licensing
SACM Reference Implementation License, Magma Protocol License, and Sovereign Logic Core Integration License for organizations building on the reference architecture.
From $15,000 / year
🛡
Enterprise White-Label Mark
Institutional partners may license the Saint Errant Sovereign Mark for proprietary certification programs subject to Saint Errant oversight and annual audit.
Negotiated
📋
Pre-Certification Assessment
Gap analysis against all five Sovereign AI Criteria. Produces a formal assessment report with specific remediation requirements prior to formal certification application.
$3,500 flat fee
Founding Declaration

Six axioms.
The institution stands.

  1. 01Sovereign AI is not a feature. It is a category.
  2. 02A system that cannot prove it governs itself does not govern itself.
  3. 03Cryptographic proof is not bureaucratic overhead — it is the evidence of sovereignty.
  4. 04Entropy is not the enemy of reliable software — it is the catalyst for resilient software.
  5. 05The WORM ledger is not a log — it is the record of a system's conscience.
  6. 06The doors of the founding guild are closed. The standard is set. The institution stands.

"Saint Errant does not wander without direction. It wanders with axioms."

Ahmad Ali Parr · Founding Director · 2026-05-20

Apply for Certification

Begin the process.

Certification applications are reviewed by the Saint Errant Standards Council. All applications receive a formal acknowledgement within five business days. Pre-certification assessment is available for organizations not yet ready for formal application.

Institution
Saint Errant — Digital Institution for Sovereign AI Governance
Founded
2026-05-20
Founding Director
Ahmad Ali Parr
Certification Enquiries
certification@sainerrant.org
Audit Services
audit@sainerrant.org
Charter
WORM-sealed · 2026-05-20 · SHA-256 anchored
Observer Portal
collectivekitty.com/observer