What is Relay

Relay is a database control plane and security layer. It solves two related problems with one model:

1. Managed databases

Provision and operate production databases without running the infrastructure yourself. Relay creates and manages PostgreSQL clusters — with backups, metrics, logical databases, users, and full lifecycle control (pause, resume, resize, restore).

You choose an engine, a provider, a region, and a machine size; Relay handles placement, networking, TLS, and day-2 operations. See Managed databases.

2. Access & governance

Put a security layer in front of the databases you already run — managed by Relay or not. Instead of handing out raw credentials, your team connects through Relay, which brokers every connection through a Relay agent and applies:

• Data masking — redact sensitive columns in query results.
• A query firewall — allow or deny queries by policy.
• Access grants — least-privilege access: members reach only the connections granted to them.
• A full audit trail — every session, and every query on connections with audit enabled, attributable to a person.

You choose how a connection gets its database password. With Store in Relay the password is encrypted at rest by the Relay server and never shown again. With a credential source (HashiCorp Vault or AWS Secrets Manager) Relay keeps only a reference and the agent fetches the password at connect time, so Relay never stores it. See Access & governance.

How they fit together

Both surfaces share the same model — organizations, workspaces, roles, and agents — so the security layer applies uniformly whether a database is Relay-managed or your own. Governance (RBAC, SSO, audit) is consistent across everything.

Next

• Why Relay — how Relay prices managed PostgreSQL and other clouds.
• Core concepts — the nouns you'll use everywhere.
• Quickstart — signup to a live connection.
• Install the CLI.