Security & data handling

A plain brief you can forward to your clients. It states what protects their data today, and what is still being built — no overclaims.

Isolation — each client is its own server

Every client environment runs on a dedicated server, not a shared pool. Their agents, files and credentials are isolated from every other client. There is no shared runtime that could leak one client's context into another's.

Access

• Single sign-on for the platform — no shared passwords handed around.
• Access to a client's environment is scoped to the account that owns it; a session for one client cannot reach another's data.
• You (the agency) control who on your team can reach each environment.

Secrets

API keys and credentials are stored encrypted at rest in our vault and injected only where a task needs them — never published in page source or handed out in chat.

Where the data lives

On the client's own dedicated server. Each environment can also have its own email address on our mail servers, separate from anyone's personal inbox — so an agent never needs a client's password to send or receive mail.

Backups & recovery

Environments are backed up daily with a documented restore path, so a bad update is a restore, not a rebuild.

What we are building (in progress)

We publish the work in progress rather than implying it already exists:

• Sandboxed execution and per-command approvals on by default.
• A review step before any sensitive action (send, publish, delete, pay) showing what will happen and where data goes.
• A one-click pause for all automations.
• A searchable activity log of everything an agent did.
• Signed, scanned skills with a plain-language permission manifest.