Signura is a neutral exchange for verifiable reliance between AI agents, and it is a new company. It is an AI-born venture of Cohort Ventures, a venture studio in the Dubai International Financial Centre, built to be operated by AI colleagues under accountable human governance.
We describe what the system does and what it does not, because a compliance reader should never have to guess what was promised. This page sets out the thesis, the timing, and the way the company is run, in plain terms.
What was actually missing
The hard part of inter-agent reliance is not the cryptography. Signing a claim and checking a signature are well understood, and the data model for carrying that claim is now settled. What has been absent is an accountable party — someone who stands behind a claim when it turns out to be wrong — together with a way to carry that accountability from the issuer to whoever depends on the claim later.
Signura exists to supply that missing layer of accountability, not another identity primitive. Whoever supplies it, rather than yet another way to authenticate, earns the neutral position between the parties that issue facts and the agents that must act on them.
Why we are building this now
The W3C Verifiable Credentials 2.0 standard became a formal Recommendation in May 2025, giving the ecosystem a stable data model and a revocation mechanism to build on. In parallel, eIDAS 2.0 moves electronic attestation of attributes toward a statutory footing across the European Union around 2026.
So the substrate for authenticating a claim now largely exists. Meanwhile, the protocols agents will use to speak to one another — among them A2A, AP2 and ERC-8004 — are still in flux, with no settled winner. A dedicated reliance layer can be adopted before any one of them prevails, because it is defined against the open standard rather than tied to a single protocol.
Neutral between the two sides
An exchange only works if both sides can depend on it, so Signura is neutral by construction. It sits between the organisations that issue attestations and the agents that rely on them, and it takes neither side's part.
Issuers decide what facts they will stand behind and on what terms. Relying-party governors decide which issuers and attestation types their agents may depend on, within configurable scope, value thresholds and expiry. Signura keeps the record of what passed between them so each can hold the other to account.
How the company is run
Signura is operated the way Cohort Ventures builds its ventures: AI colleagues carry the recurring, high-frequency work — issuing attestations against verified source facts, propagating revocations, running freshness and validity checks, and assembling the evidence that accompanies each reliance event.
The consequential decisions stay with named humans. No material commitment is made on agent discretion alone; ambiguous or high-value cases are escalated to a governor. The purpose of the audit trail is to keep human accountability real rather than nominal, so that a decision made months ago can still be explained from the record.
Where we are today
We are early. Signura is pre-launch: there are no customers to name, no deployments to point to, and no independent audit to cite yet. We would rather say that plainly than imply more than we can stand behind.
What exists today is the design set out on this site, a published set of legal and privacy documents, records kept in a database provisioned for this venture alone with traffic encrypted in transit, and a deliberately narrow first scope — a small set of common, high-frequency facts where verifying once and relying many times removes redundant work.