Security

Least privilege, isolation and auditability by default. We describe only controls that are actually in place.

Accounts

Passwords are hashed with scrypt. Sessions are server-side, httpOnly and expire; you can sign out of all devices. Repeated failed sign-ins lock the account temporarily. Two-factor login with an authenticator app is available and can be required for administrators and other users with access to contracts, datasets or payouts.

Administration

Internal access is split into roles (finance, data operations, compliance, contracts, security). High-value payouts require a second administrator. Privileged actions are logged and alerting is in place for unusual events. An emergency lockdown can suspend all non-administrative access.

Payout protection

Changing payout bank details requires re-authentication, notifies all company owners, resets verification and pauses payouts for a cooling-off period.

Access

Read-only scopes where the provider supports them. Credentials are encrypted with AES-256-GCM using an environment-managed key and never leave the server.

Isolation

Each supplier's data is isolated with PostgreSQL row-level security in addition to server-side authorization on every action.

Delivery

AI developers only receive licensed versions via short-lived signed links or scoped, expiring API keys. Every access is logged.

Operations

Append-only audit log and ledger, rate limiting, content security policy and other security headers, webhook SSRF protection, and automated tests for tenant isolation, access control and payout logic. We have not completed an independent penetration test or certification and do not claim any.

Report a vulnerability

See our responsible disclosure page or /.well-known/security.txt.

See exactly what AI buyers receive

Switch between an original business record and the proposed licensing dataset, with every excluded, generalized and reviewed field marked. Synthetic examples.

See How Your Data Is Protected