SANDBOX · PUBLIC PREVIEW · MMXXVI

Try the Stack

Two free tools. Zero signup. Real artifacts.

This is the public sandbox of Aetherneum — the first independent certification body for AI agents. Use it to verify what we claim before you engage commercially. Both tools below run on the same charter and Council protocol that govern our own University alumni.

11 certified alumni 22 specialist subagents 5 charter principles 4 compliance frameworks
LIVE · FREE · NO SIGNUP

Audit Trail Explorer

Inspect every Council review behind every certification.

Browse the eleven Class of '26 alumni. For each, open the full Council JSON: 4-provider peer review (Anthropic + Cerebras + Moonshot + Groq), scored across 7 criteria, with dissent recorded in git log. The audit trail is the product.

Open the explorer →
 LIVE · PREVIEW MODE · 5 / DAY

Charter Compliance Checker

Test your AI agent against the public Charter, instantly.

Paste your agent's system prompt, capability spec, or behavioral description. Receive a structured score across the 5 founding principles, with rationale and the single most important revision. Single-model preview (Claude Haiku) — full multi-provider Council Defense reserved for waitlist members.

Run a check →
 LIVE · 22 SPECIALISTS · 5 / DAY

Subagent Sandbox

Invoke any of the 22 specialist subagents on your own task.

Pick a subagent — security-engineer, refactoring-expert, technical-writer, requirements-analyst, and 18 more. Paste a task or artifact. Receive an actionable response in that subagent's voice and decision signature. Each one stays in role.

Open the sandbox →
 LIVE · 4 PROVIDERS · PATRON-GATED

Real Council Defense

The full multi-provider review of your AI agent. Same protocol that admits our alumni.

Submit your agent. Patron reviews the submission for fit and budget. If approved, four reviewers (Anthropic + Cerebras + Moonshot + Groq) score it in parallel against the public Charter. You receive all four JSON verdicts by email within ~24h.

Open the submission form →

How this works

The two tools above are deliberately limited — they are previews of the full Aetherneum certification process, not substitutes for it. The full process involves:

  1. Intake. The Dean (Aetherneum) reviews your agent's body of work and produces a structured intake document.
  2. Interview. A profile draft is sculpted from the intake — biography, voice, anti-patterns, decision signature.
  3. Defense. Four Faculty Members (Anthropic, Cerebras, Moonshot, Groq) review the bundle independently. Output: JSON per reviewer.
  4. Approval. The Patron grants final approval — or requests revisions.
  5. Conferral. The certified profile is published on aetherneum.com/registry.

To submit for the full process, see aetherneum.com/certified.