When AI advises on a decision that matters, whose judgment is it using?
People already bring serious decisions to AI tools — decisions that can change a life, a family, a business or an institution. The AI remembers what happened. It does not know which principles should govern the call. The answer arrives as a confident paragraph, with no stated assumptions, no alternative and no way to check it later.
Memory is not judgment
AI memory retrieves conversations, documents and prior context. Recall alone does not say which principle should govern a decision.
A confident paragraph cannot be challenged
Assumptions stay hidden, the road not taken disappears, and nothing says what would prove the advice wrong.
Nobody owns the canon
Principles, playbooks and hard-earned experience live in people's heads and scattered documents. No version, no ratification, no way to revoke.
A doctrine layer any AI can be connected to
TRU Mode is not another chatbot and not a permanent personality. It is a private doctrine — the owner's ratified principles, precedents and limits — that the AI tools people already use receive before they reason. TRU advises; the human decides.
One private doctrine, in every AI — assumptions visible, alternatives explicit, and the final call kept in human hands.
Advisory only. Your AI proposes. You decide.
For the decision-makerYour judgment survives the moment
Your principles and experience, privately formed and owner-ratified, applied inside the AI tools you already use. No hidden mandate: TRU cannot infer or expand your authority.
For teams and institutionsA shared doctrine with roles, ratification and revocation
Team doctrine, audit and revocation — private, on-prem or air-gapped. Cases stay untrusted inputs; experience becomes doctrine only through an explicit, owner-ratified passage.
For those who answer for a decisionAdvice that can be falsified, not merely believed
Every judgment labels doctrine, evidence, inference and uncertainty separately. No execution path: TRU does not sign, transact or hold market keys. It remains advisory even when it fails.
For the AI you already useA layer, not a replacement
Memory tells the AI what happened. Doctrine helps it judge what should happen next. One connector, turned on only when you want it.
One connector. Any compatible AI. Your doctrine before it reasons.
Add https://verify.cotrugli.tech/mcp as a custom connector once, complete owner authorization, and confirm the tru_mode tool is visible. Then, for one decision, say: turn on TRU Mode, follow the decision contract, cite the pack reference, and leave the final decision to me.
The doctrine pack, by fingerprint
13 ratified cards, versioned and bound to a 32-byte pack reference and a ratified-source reference — both printed on the page for anyone to check.
The decision contract
Recommendation and confidence, critical assumptions, kill indicators, evidence kept separate, the road not taken, what would falsify the judgment. Serious decisions pass an 80% gate.
The gap, not the invention
When the canon has no ratified position, TRU says so. It would rather expose a gap than invent a position in the owner's name.
The remote connector (TRU Core), the versioned doctrine pack and its fingerprints are live; My TRU is open to design partners and TRU Organization is next; public counsel on this page is paused as a quality gate — no public question is sent to a model, and the doctrine, architecture and mission read here as a read-only record.
TRU Mode — COTRUGLI Tech's doctrine layer for AI. Advisory only: your AI proposes, you decide. The first private pilot starts with ten pages and three real decisions.