Every agent action carries proof of who authorized it, the limits it ran under, and AI guarding with an E-STOP that fails closed.
E-STOP is a current platform capability. Revocation, policy deny, and fail-closed gates stop an action that is outside authority.
Payment authorization is not legitimate decision-making authority.
A portable, independently verifiable claim that an agent’s specific action falls within a valid chain of delegated authority, satisfies applicable constraints, and has an attributable outcome.
When an agent acts on your behalf, don't get left holding the bag.
Enterprise software assumes identity, permissions, contracts, and logs are enough. Autonomous agents act across organizational boundaries, where those assumptions stop being a shared system of record.
Existing assumption
Identity establishes who you are
Replacement
Cryptographic identity establishes who or what is acting
Existing assumption
Permissions establish what you can access
Replacement
Delegated authority establishes what you are entitled to decide and do
Existing assumption
Applications own their workflows
Replacement
Independent agents coordinate across organizational boundaries
Existing assumption
Logs establish accountability
Replacement
Verifiable receipts connect authority, evidence, decisions, and outcomes
EXOCHAIN does not replace MCP, A2A, x402, or AP2. An agent can be allowed to pay and still lack the authority to choose the vendor, the evidence for that choice, or a receipt a counterparty can verify on its own.
Connects AI systems to tools and resources.
Lets agents collaborate across systems.
Carries machine-native payments.
Carries signed mandates and receipts for agent-initiated payments.
The intended experience is two calls: verify that the action is within delegated authority, then attest the outcome. The current SDK is the 0.2.7 client in the repository. The snippet below is the target shape.
const authority = await exochain.verify({
agent: agentIdentity,
action: proposedAction,
mandate: delegatedAuthority,
context: decisionEvidence
});
if (!authority.allowed) {
throw new Error("Not authorized");
}
const receipt = await exochain.attest({
authority,
decision: proposedAction,
outcome: executionResult
});EXOCHAIN should be useful to companies it did not create and does not control. CrossChecked is the first commercial wedge, AVC is the distribution engine, and the AI-SDLC Institute keeps education and assessment independent.
The shared trust fabric: open protocols, SDKs, and enterprise services other companies can use without being owned by EXOCHAIN.
Shows who authorized an AI-assisted decision, which evidence and constraints applied, and what occurred afterward. It is an adjacent product, currently a prototype, and it does not itself enforce the constitutional kernel.
Incubation and venture formation for companies built on the fabric. Distinct from the protocol’s Autonomous Volition Credential.
Training and assessment that stay independent of the products in the commercial network. Assessment here is education, not coverage.
CyberMedica, LegalDyne, LiveSafe, and Decision.Forum are reference implementations and candidates for future independent businesses. They are not six products this site is launching at once.
Offers are scoped in conversation. This site does not publish a rate card.
A fixed-scope design-partner pilot or deployment for one constrained workflow, including architecture and integration into the systems you already run.
Licensed access to the EXOCHAIN control plane, hosted verification, and enterprise integration. The platform is the recurring product.
Ongoing administration, governance engineering, monitoring, remediation, and platform optimization, purchased as an outcome.
Bring one constrained, cross-organizational workflow — a procurement decision or an executive approval, for example — where two parties need to verify legitimate authority without sharing an internal system. Developers can start from the SDK in the public repository.