You carry the token from the front desk to the physician to accounting. You install SIUI on every computer. You keep track of "which machine has the certificate today".
ZenAgent · automatic CNAS claims
One CNAS token.
The whole clinic reports.
The token for CNAS (Romania's national health insurance house) and the card reader sit on one station, connected once. From there, every computer in the clinic reports to CNAS straight from the browser — ZenAgent routes each operation to the right station, automatically. No token carried from desk to desk, no SIUI (the legacy desktop reporting app) installed on every machine.
The request starts from any station. ZenRIS routes it to the station that holds the eligible token. The PIN is entered only there. Nobody moves anything.
Centralized in the RIS, not scattered across stations
The RIS is the brain. It decides which CNAS operation is needed, for which contract and with which certificate, then sends agents typed commands over a secure channel. The local agent holds no authority — which tenant, which station and which permissions are all set by the server, from enrollment and identity.
Every command has a deadline, a correlation ID and is bound to a specific device. Nothing runs on trust coming from the station.
- Typed commands — not free-form scripts; every operation has a clear contract.
- Server-derived authority — the station doesn't declare its own permissions.
- Complete audit — every CNAS operation is logged: who, what, when.
The agent installs once, where the hardware is
ZenAgent runs in the background, as a service, on the station that holds the CNAS token and/or the card reader. It installs per-machine, starts inert and activates only after enrollment. Updates are signed, verified, with automatic rollback if anything is unhealthy.
The operator never opens or configures an agent application day to day. They work only from the RIS — the agent stays invisible until it's needed.
- Per-machine install, inert until the station is enrolled.
- Signed updates, verified by manifest and hash, with health-gated rollback.
- Zero daily setup for the operator — everything from the RIS.
This is the difference
The token works for the whole clinic — without anyone moving it.
When a station without a token requests a CNAS operation, the RIS routes it automatically to the station that holds the eligible token for that operator, tax ID and contract. Deterministic, with no fallback between certificates: if the token is missing, expired or ambiguous, the system stops safely (fail-closed) — it never guesses.
The card reader is always local to the station you work from. The token can execute for another station, but only through an explicitly authorized route — and the PIN stays on the token's station.
- Routing by capability, not by whoever happens to have the token plugged in.
- Fail-closed — on any uncertainty it stops, it doesn't invent a signature.
- The clinical flow continues even if the token or CNAS is momentarily unavailable.
The keys stay on the token. Always.
ZenRIS never stores P12/PFX certificates, PINs or private keys. Security isn't a promise — it's a consequence of the architecture.
- The private key stays on the hardware token — ZenRIS neither sees it nor copies it.
- The PIN is strictly local, on the token's station, even if the request came from another desk.
- Non-exportable station identity, with a cryptographic browser–station binding.
- Certificate bound to operator–tax ID–contract, with a full audit of every operation.
One token. One station.
The whole clinic reports.
The token sits on one station, connected once. Any computer reports from the browser. Routing is automatic, the PIN stays local, everything is audited.
The same agent architecture extends to fiscal devices (cash registers) — through a dedicated adapter, verified on a real device. Clearly marked as roadmap, not a feature of today.
CNAS claims become a non-event.
Bring your token, your contract and your compliance questions to the demo. We'll show you how a single ZenAgent station covers the whole clinic — and how everything stays audited, fail-closed and with the keys on hardware.