know your agent.

every agent identified. every mandate sealed before the agent moves. every action chained to the authority that allowed it.

the problem

kyc exists for the humans. nothing exists for the agents. ai agents already work inside your stack on borrowed credentials, and when one of them acts the first questions are which agent, acting for whom, with what authority.

syenis that proof. authority is sealed before the agent moves, not reconstructed after. every mandate and action is signed with ed25519, chained with sha-256, and anchored daily to digicert’s rfc 3161 trusted timestamp authority.

what each record proves
  1. 01who

    which agent acted, and who it acted for.

  2. 02authority

    the mandate it was given, before it acted.

  3. 03action

    what it actually did, and when.

seal before action
before it acts

the mandate is sealed and signed. authority exists as proof, not a config line.

while it acts

every action chains to the mandate that allowed it.

when it is challenged

show the record. it verifies without you in the room.

when it drifts

an action outside the mandate fails verification against it.

independent verification

proof that stands on its own.

your auditor checks the entire chain with openssl. no access to your systems. no access to ours.

ed25519 signatures·sha-256 chain·rfc 3161 anchoring·openssl verification
where teams use it

teams answering ai governance questions under the eu ai act, nist ai rmf, or internal audit. see the governed workflows on the use cases page.