Third-Party Custody vs. Self-Custody Architecture

Third-party custody is an arrangement in which a third-party operator holds and controls the cryptographic keys on behalf of a user, routing every transaction through its own infrastructure.

Third-Party Custody Architecture

• Custodian-Managed Routing: every request passes through the custodian's own infrastructure.
• Custodian-Held Keys: keys and balances are held and managed by the custodian.
• Third-Party Trust Dependency: users depend on the operator's own controls and availability.

Self-Custody Architecture

• Transaction Request: the user initiates the action directly.
• Key-Based Authorization: the user's own key authorizes the transaction directly.
• Transaction Signing & Broadcast: the user, not an intermediary, finalizes the transaction.

That shift — from an operator's infrastructure to keys the user actually holds — is what people mean when they say self-custody.

See the related breakdown of distributed key signing for the cryptographic mechanics behind this architecture.