Trust Center

The component that acts on trust decisions runs inside your environment, and it alone holds the credentials for your enforcement points. We never have them, and there is no path from our side into your infrastructure. Nothing your people do passes through us: no proxy, no tunnel, no point of presence. What travels between us is short enough to list, and checkable: agent check-ins, the exchange between your trust control server and the service that scores attestations, token activation email, and signed agent updates.

Answering “is this still the device we trusted?” needs the device’s attestations, the token holder’s email address, and a record of trust and administrative changes. We hold nothing else. In particular, we hold none of the following:

• ✕Names, phone numbers, addresses, or HR records
• ✕Browsing history, files, or application content
• ✕Screenshots or keystrokes
• ✕Running processes
• ✕Your organizational device, account, or system structure

Trust is also scoped per token, not per machine. A device can carry a work token and a contractor token side by side, and what one observes never reaches the trust decision of the other. Each organization sees only the access it issued.

Nobody, with one exception that exists so we can bill and renew you: our commercial systems hold account, contract, and entitlement records. They cannot reach your device data, trust history, or audit records, and destructive actions such as deleting an account require a second person’s approval.

An agent that runs on every machine in an organization should be judged on its limits. It runs without administrative rights, and it cannot:

• ✕Raise its own trust. Scoring happens on our side; the agent reports what it observes and applies what it is told.
• ✕See your traffic. There is no proxy, no VPN client, no filter, no inline hook of any kind.
• ✕Touch anything on the device beyond its own state. It is not a management agent.
• ✕Issue or revoke a token, including for another user of the same machine.
• ✕Enroll itself. A device cannot admit itself into trust; an administrator provisions the person, or it does not happen.

Everything in transit is encrypted, every agent message is signed and verified, device secrets live in that device’s security hardware, and updates are signed. We use standard cryptography only, no primitives of our own. The service runs in the United States; if your obligations require processing somewhere specific, raise it during scoping. Retention is set by your license, and administrators can revoke any device at any time without asking us.

If Veribound is ever unreachable, your traffic is unaffected: your edge keeps enforcing the last state everyone agreed on, trusted stays trusted, and untrusted stays out. What we ask of your systems is equally narrow, and it is described on the Integrations page.

A public page can responsibly carry only so much about how a security product is built: detail that helps your review also helps an attacker, so the two are kept apart. The full package is shared with prospects and customers under NDA.

Tell us what your review needs and we will assemble that material for your team.