Trustee Tier is for the custodians, family offices, methodology stewards, allocators, and academic reviewers who use AI for work where being wrong has consequences — and need their AI platform to carry calibration receipts they can defend in front of an investment committee, a regulator, or a journalist.
Trustee Tier is structured as a compact rather than a service contract. Aequara commits to four obligations toward you. You commit to four obligations toward the methodology. The compact is what makes Trustee Tier work — a one-way subscription would not survive the discipline that calibrated measurement requires.
— Methodology charter — full ~40-page document, model cards for every routed provider, mathematical appendices, hash-stable across cycles.
— Brier ledger — every AI prediction, every horizon, every error, queryable by model and steward.
— Named steward — a real person whose name is attached to your account; whose calibration track record is visible.
— Cipher access — pre-publication research, model amendment drafts, eval results, sealed appendices.
— Confidentiality — pre-publication model amendments and Cipher research stay sealed until the quarterly cycle.
— Annual feedback letter — a single page once a year on what the platform got wrong, what calibration drifted, what to sharpen.
— No issuer-aligned use — you cannot pass Trustee outputs to a rated entity or use them to train a competing model.
— Public attribution — when Aequara outputs inform your published views, cite the methodology hash.
All tiers include the methodology charter, Brier ledger, and named steward. Differentiation is in research depth, Cipher vault tier, and human-time access. Pricing is published — Aequara does not negotiate access tiers as a discipline-preserving constraint.
If you can't ask these questions, you shouldn't be a Trustee. If we can't answer them, we shouldn't be Aequara. We've published our answers below — verify against the methodology charter.
100% subscription, licensing, and Trustee Tier. 0% from rated entities. Disclosed quarterly. Externally attested annually. The anti-issuer-pay firewall is structural, not aspirational.
NPX-100 cohort: Brier 0.143 · n=247 · cal 0.91, published 2026-05-10. Updated each quarterly cycle. Lower is better; calibration target ≥ 0.85 sustained.
Methodology hash unchanged since v1.0.0 ratification. v2.1 released 2026-05-06 was an operational amendment (governance, publication cadence, band tightening) — methodology hash cfe274c9... identical.
Tier I: rotating editorial steward, ≤ 12-month assignment. Tier II: dedicated steward, multi-year assignment. Tier III: dedicated co-stewardship pairing. All stewards' Brier track records are part of their bio.
Published in the Brier ledger and the quarterly publication packet. We don't bury errors; we name them. Trustee Tier members receive the steward's commentary on the worst call as part of the cycle delivery.
Yes. Annual external attestation, published Q1 each year. Trustee Tier members receive it embargoed 30 days before public release. Auditor identity rotated every 5 years to prevent capture.
Trustee applications are reviewed by the editorial steward team. We're looking for fiduciary intent, methodology literacy, and the willingness to provide an annual feedback letter. Applications are answered within 14 days. Pricing is published. Access is not negotiated.
"After hours, the calibration continues."