Connections

A connection is a structured, reusable object that holds everything a workflow needs to connect to an external system such as a database, an object store, or a SaaS API. Instead of scattering hostnames, ports, and credentials across many nodes, you define a connection once and reference it.

Connection types

Every connection has a connection type that defines its shape as a schema. For example, a postgresql connection type requires host, port, user, dbname, and a password secret field. Defining the type lets the editor validate connections and offer autocomplete.

Field kind	Stored as
Plain config (host, port, region)	Variable — readable
Credentials (password, token)	Secret — encrypted, write-only

Defining a connection

Connections live at a path. Scaffold one locally, fill in its spec, then deploy:

orvanta connection new f/folder/prod_db
# edit the generated spec, then:
orvanta sync push

Inspect or list connections with orvanta connection get <path> and orvanta connection list.

Using a connection in a workflow

Reference the connection by its path from a script or flow step; Orvanta resolves the fields (decrypting any secret fields) at execution time.

This keeps connection details consistent across every workflow that talks to the same system, and means rotating a credential is a single update.

Related

• Variables and secrets: The building blocks connections are made of.
• Workflows: Where connections are referenced.
• Roles and permissions: Restrict who can use a connection.