Human review at every step
The loop pauses before the agent acts and after it verifies the result. Either point can route to a person, then the run resumes.
Human review at every step, an append-only audit trail of every action, encryption with your own key, and tenant isolation enforced at the database layer, with single-tenant deployment in your own AWS or Azure environment available.
SOC 2 Type II & HIPAA in progress Trust Center →Every run is observable, interruptible, and reviewable by your people.
The loop pauses before the agent acts and after it verifies the result. Either point can route to a person, then the run resumes.
Every run produces a structured event trace: the step, the action, the outcome, and any retry reasoning.
Operators see what the agent is doing and what needs approval, right in the browser.
The deployment boundary is yours to choose, and so is the model endpoint the agent talks to.
Sensitive fields are encrypted at rest with a key you supply. Without that key, the deployment refuses to start.
The agent sends page context to whichever model endpoint you configure. Point it at a model you host and that context never leaves your infrastructure.
Multi-tenant by default in our cloud. Deploy ModelNex single-tenant in your own AWS or Azure environment when your security review calls for it.
Every action is written to an audit log that cannot be edited or deleted in place, retained for seven years by default.
Customer data is never shared across tenants; isolation is enforced at the database layer.
Honest status, not badge collecting: our SOC 2 Type II and HIPAA programs are in progress and monitored continuously with Vanta, and the Trust Center shows where each one stands.
Where each program stands, the controls in place today, and the evidence behind them, updated as it changes rather than once a year.
Open the Trust CenterOur SOC 2 Type II program is in progress and monitored continuously. Control-by-control status is live in the Trust Center.
Our HIPAA program is in progress. In place today: sensitive data encrypted at rest with your key, role-scoped access, and every action in the append-only audit trail.
A security overview document is available on request.
Data is encrypted in transit with TLS and at rest with AES-256.
We never use customer data to train models. Model calls can be routed to endpoints you host, so data stays inside your boundary.
Retention periods are agreed per contract, and customer data is exported or deleted at termination.
We keep the list short. A current subprocessor list is available on request.