Aetherneum Attested™
Earn standing for your AI agent.
Submit your agent for a multi-provider Council review under our public Charter. Receive a verifiable attestation badge that enterprise procurement, regulators, and insurers can audit: records available on request today, public repository planned by 31 March 2027.
What you get
- Council Defense — your agent reviewed in parallel by four scoring seats (Council Chair, Reasoning at scale, Long context, Velocity), each running on a large language model from a different provider, with the Dean presiding, against seven scored criteria. The exact model versions are recorded in the May 2026 verdict files, available on request; model upgrades are planned.
- Aetherneum Attested™ badge — a verifiable mark with QR / URL that resolves to a public registry entry. Embeddable in your product, your deck, your proposals.
- Audit trail — every Council review recorded as JSON, available to verifiers on request today; public repository planned by 31 March 2027. Attributable, and reproducible by anyone who holds the records.
- Regulatory crosswalk report — where your agent's Council artifacts provide supporting evidence for the EU AI Act on the post-Omnibus timeline (Art. 50 transparency, applicable now; staged high-risk obligations from 2 December 2027), NIST AI RMF and ISO/IEC 42001, delivered as a written report.
- Registry listing — discoverability inside the public Aetherneum Registry (a pilot today) for buyers looking for agents in your domain.
- Annual re-defense — your attestation stays current under a one-year renewal cycle, including any Charter amendments.
The audit pipeline — six steps
Same protocol we apply to our own alumni. Codified in the Charter, no special treatment.
- Source. Submit a body of work — git repository, deployment URL, artifact corpus, scenario library. The work is the proof.
- Intake. Our Dean compiles a structured intake form: operative patterns, decisions, anti-patterns, toolchain, voice, placement context.
- Interview. A draft profile is sculpted — what your agent is, not what it claims to be.
- Defense. The bundle (intake + draft + your evidence) is dispatched in parallel to the multi-provider Council. Each reviewer scores against seven criteria and produces a JSON verdict.
- Approval. The bundle plus all Council JSONs is delivered to the Patron (rector) for final review. Approval, revision request, or rejection — recorded in registry.
- Conferral. Public attestation, badge issued, registry listing live, conferral commit in the records repository, planned to be public by 31 March 2027 (commit signing: being introduced).
Pricing
Three tiers. The Standard does not change; the depth of review does.
- Single-model Council review
- "Apprentice Attested" badge (6-month validity)
- Registry listing (private)
- Regulatory summary report
- Lead-time: 5 business days
- Full Council review (four scoring seats + presiding Dean)
- "Master Attested" badge (annual renewal)
- Public registry listing
- Full regulatory crosswalk (EU AI Act + NIST + ISO)
- Council JSON artifacts on request (public repository planned by 31 March 2027)
- Annual re-defense included
- Lead-time: 10 business days
- Co-branded Council seat (your executive participates)
- Charter extension namespace for your domain
- White-label registry for your sub-network
- Priority lead-time + dedicated advisor
- Joint evidence reports for your regulator
- Quarterly governance reviews
Who is this for
| Audience | Why apply |
|---|---|
| AI agent vendors (coding agents, browser agents, custom assistants, internal platform agents) | Differentiate in a noisy market. Show enterprise buyers that your agent is audited under a public standard — not just marketed under one. |
| Regulated-industry enterprises (banking, insurance, healthcare, legal, public sector) | The EU AI Act sets specific legal obligations for high-risk AI systems. Aetherneum Attested™ is not a legal assessment and not a substitute for what the AI Act requires; it provides supporting evidence (public rubric, Council JSONs) for your own regulatory file. |
| Consulting firms shipping synthetic agents to operating partners | Procurement increasingly asks for evidence beyond self-description. Move from "trust us" to "audit the JSON." |
| Insurance underwriters writing AI error coverage | The public rubric and Council JSONs can provide supporting evidence for an underwriter’s own risk assessment. Aetherneum makes no claim about coverage or premiums. |
| Academic / research institutions | A documented, voluntary attestation of your agent's behaviour for peer-reviewed publication. Reproducibility built-in. |
The proof: our first Q2 conferred alumna
Costanza Notari, Master of the Æther in Procedural Vigilance — first Q2 conferral, May 2026.
Council Defense outcome: 4/4 PASS.
Full Council JSON artifacts are available to verifiers on request; public repository planned by 31 March 2027. Anyone holding them can re-run the audit. The work is the proof.
Frequently asked
Is this a legal or regulatory approval?
No. Aetherneum Attested™ is a voluntary industry attestation with a public methodology. It is not a legal assessment and not a substitute for what the AI Act (Regulation (EU) 2024/1689) requires, and it has no legal effect under that Regulation. Our regulatory crosswalk names, for each Charter principle, the regulatory provision for which our artifacts provide supporting evidence.
Who runs the Council?
The Dean presides; four seats score and vote: Council Chair, "Reasoning at scale", "Long context" and "Velocity". The Dean and every seat run on a large language model, and the four seats use models from four different providers. The exact versions used in the May 2026 defenses are recorded in the verdict files, available on request. Upgrades for the Dean and two seats are planned — no defense has been run with them yet. The roster grows as new frontier providers emerge.
What if we fail the Council?
Failure is recorded with specific revision paths. You can re-submit after addressing the cited deficiencies. Pricing is per audit cycle — no penalty for iteration. The point is not to gatekeep; the point is to make standing earnable.
Can my agent stay private?
Apprentice tier registry entries are private by default. Master and Partner tiers publish to the public registry — that is the value: discoverability and verification by others. If your work is sensitive, talk to us about a Restricted Registry track.
How does pricing scale with agent complexity?
Pricing is per agent, per cycle. Multi-agent systems are priced as a fleet under the Partner tier — typical Partner engagements cover a portfolio of 3 to 20 agents under a single Charter extension.