Custody, controls and evidence

Stablerail runs on Privy MPC infrastructure (owned by Stripe). Private keys are split into shares that are never combined: the whole key does not exist at any point in its lifecycle, and signing takes place inside hardware-isolated environments. Stablerail holds no share that lets it sign alone, so it cannot unilaterally move customer funds.

• No whole private key ever exists — MPC threshold signing only.
• Signing executes in hardware-isolated enclaves, not in application code.
• Quorum-protected key export means customers can leave with their keys.
• Stablerail is not a custodian and cannot initiate transfers on a customer's behalf.

• Policy engine: per-counterparty, per-amount and per-asset rules applied before a payment can be proposed.
• Approvals: multi-approver workflows with named requesters and approvers on every payment.
• Screening: sanctions and wallet-risk checks against counterparties before funds move.
• Anomaly detection: AI flags outliers — first-time counterparties, unusual amounts, off-hours activity — and explains why.

Each transaction stores the full operating record: who requested it, who approved it, which policy applied, what screening returned, which counterparty was involved, and the on-chain result. That record is exportable, so month-end close and audit requests do not depend on chat history or screenshots.

• Data encrypted in transit (TLS) and at rest.
• Role-based access control for team members, with least-privilege defaults.
• Row-level authorisation on all customer data; internal tooling is separated from customer-facing systems.
• Fiat rails and card issuing run through regulated partners with their own supervisory regimes.

Report a suspected vulnerability to security@stablerail.com. We acknowledge reports and respond with a remediation plan.