Programmable trust — without programmable control. Trust State is a layer that lets a verified fact travel between platforms while no platform, administrator, or model holds an override, because the architecture has none to hold. It records what is attested and renders it legible. It is a scale, not a judge. Open now for reading, evaluation, and licensing conversations.
Nothing crosses without its instrument.Every platform re-verifies the same people from scratch, because the systems that can share trust also share control — the power to freeze an account, force a transfer, override a state. So nobody shares. Trust State separates the two: the fact moves, the control does not.
Portable identity has always come bundled with an administrative surface — someone, somewhere, can reach in. Platforms that can’t accept that surface are forced to silo, and re-verify everyone themselves.
The layer holds faithful registration of attested facts and nothing more. Meaning is applied at the edges, by the parties who bear the consequences. There is no control surface to capture, because none is built.
Comparable programmable-compliance designs include a control box — freeze, forced transfer, override. This layer deliberately omits it. That single absence is the product.
A platform gains trust that carries across the network while still setting its own thresholds — no central controller sits above the network, and it hosts no one else’s policy engine. Interoperate on facts; keep your own judgment.
How can trust become programmable without making control programmable?
The Global Layer One initiative — with contributors from the IMF, Banque de France, Kinexys by J.P. Morgan, and the Monetary Authority of Singapore — published the reference model for programmable compliance. Its own risk section names two problems the model leaves open. Both filings predate it. GL1 is evidence of the convergence, not its cause.
Their model wires in freeze, forced transfer, reversal. Their own paper warns it becomes a prime target: outside attack, inside abuse alike. Their answer to the concentration risk is more governance over the override. Ours is not having one.
Their paper concedes a point-in-time check quietly lapses. Here the state tracks its sources continuously — by construction, not by policy.
The most demanding room in finance described the shape of the answer. The answer was already on record.
Find the override. Then choose the architecture that doesn’t have one.
Start with the public thesis, or come straight to the principals if you’re ready to evaluate. Either way, every disclosure is preceded by the instrument that governs it.
Access is recorded. Submitting registers your acknowledgment of the terms and releases the public whitepaper to your business email.
This acknowledgment covers receipt of the public whitepaper only. It grants no license and no rights in either principal’s patent-pending IP. The whitepaper states the architecture; its structure and the commercial and ownership terms open only at evaluation, under a signed mutual NDA.
Four stages. At each one, more of the architecture crosses — and a stronger instrument governs it first. The order does not invert. You are only ever one step in; the rest is here so you can see the whole path before you take it.
The thesis, the frame, the posture — enough to decide whether to pursue, nothing that enables a build.
Crosses: concept & commercial frame. On record: requester identity + IP acknowledgment.
Layer-level materials for diligence — reviewed under a mutual NDA governed by the principals’ chosen jurisdiction, with non-use, non-circumvention, and a defined confidentiality term. This is where the architecture’s structure, the principals, and the commercial and ownership terms are disclosed. Evaluation is not co-development: nothing is authored jointly, and no rights arise from review.
Crosses: principals, structure & evaluation detail. Governs: non-use, non-circumvention, defined term.
Two tracks, set by rights tier, not by scope or depth. Operate — you hold a position in the layer and may extend and sub-license beneath it. Use — a terminal grant for your own purposes. Tier is fixed at signing and does not upgrade by depth or later request; pricing is set by the layer’s defined mechanism; licenses grant use of the layer — never ownership of either principal’s estate.
Crosses: the right to operate on / use the layer. Governs: tier, floors, term, pricing deference.
The conditions that hold for the life of any license — fixed estates, indivisible principals, and the architectural posture that determines what the layer will and will not host.
Crosses: nothing further. These terms bound every stage above.
A counterparty should know these before evaluation, not discover them at signing. They are why the layer cannot be captured, split, or re-hosted.
Trust State is co-held by its principals. No engagement, evaluation, or co-authored work transfers or creates rights in either principal’s estate. Pre-existing IP does not become a deliverable.
Engaging one principal grants no rights in — and does not circumvent — the other. The layer is engaged through this sequence, not around either party.
The layer records attested facts and renders them legible. It does not judge, does not decide, and will not host a counterparty’s policy engine, middleware floor, or adjudication layer. There is no attachment point for one. Meaning is applied at the edges, by the parties who bear the consequences — the layer holds only fidelity.
Who holds what, in which entities, under which filings — disclosed at evaluation, under NDA, not before. This page asserts that the layer is protected; it does not map how.
Trust State is its own front door — not routed through either inventor’s company. Bring your evaluation to the layer directly. Engaging it engages both inventors jointly, never one around the other. Who holds what, in which entities, and on what terms opens once you’re inside, under a signed NDA.