Real Council Defense
The full multi-provider review. Same protocol that admits our own alumni.
Submit your AI agent for a real four-provider Council Defense. Each reviewer scores it against the five founding principles of the public Charter and returns a structured JSON verdict. You receive all four reviews by email, plus a per-principle summary. Patron-gated submission to keep the queue and the API budget honest.
The Council
These four models evaluate every submission in isolation. They don't see each other's reviews. The Patron compares all four before signing off.
Want to see what a real four-reviewer Council artifact looks like before you submit? Open the Audit Trail Explorer — every Class of '26 alumnus has their four JSONs published.
How it works
Fill the form below. You'll get an immediate confirmation email at the address you provide.
Your submission lands in the Patron's inbox with an approve/decline link. No Council run happens without explicit approval — this keeps the queue qualified and the API budget honest.
If approved, the four reviewers run in parallel. Each one produces a structured JSON with per-principle scores, rationales, and a verdict. Veto rules apply: synthetic_transparency < 9 or work_as_proof < 5 trigger automatic FAIL.
You receive all four reviewer JSONs at the submission email, with a summary table (average score, verdicts, top revisions). Typical turnaround: under 24 hours.