Bruin CLI

Appearance

Connections ​

Connections are named configurations that let Bruin authenticate and talk to external systems: data sources, destinations, and any other platform a pipeline or agent depends on.

Locally, you define connections in .bruin.yml. That file should be gitignored since it holds secrets and credentials. In Bruin Cloud, you configure the same connections through the UI. The names need to match the ones referenced in your pipelines, and they are encrypted at rest using HashiCorp Vault as the secrets backend.

Add a connection ​

1. Spot what is missing ​

Open a pipeline. On the right-hand side, Bruin Cloud lists the connections it uses. Any connection referenced in your pipelines but not configured in Bruin Cloud is flagged so you know what to fix.

You can also work from a clean slate: open the team menu in the top bar and click Connections (under the Manage Team section).

2. Open Connections and click "New connection" ​

From the team menu, click Connections and then New connection.

3. Pick the connection type ​

Select the type that matches your data platform (BigQuery, Postgres, Snowflake, Databricks, etc.) or pick a generic secret for API keys, tokens, and other text-based credentials. See Connections overview and the platforms section for the full list of supported types and per-platform config.

4. Match the name ​

The connection name must match the one referenced in your pipeline exactly. If they do not match, the pipeline will keep flagging the connection as missing.

5. Fill in the details ​

Enter the rest of the connection details. Each platform asks for different fields: access key, host, project ID, username and password, and so on.

6. Create and validate ​

Click Create. For most data platforms, Bruin Cloud validates the connection by running a SELECT 1 against the database. If you do not want that, choose Create Without Validation.

7. Confirm it is active ​

Go back to your pipeline. The connection should appear in the right-hand list without the missing flag.

Generic secrets ​

For text-based credentials that are not tied to a specific data platform (API keys, OAuth tokens, third-party secrets), use the generic secret connection type. They work the same way as named connections in .bruin.yml: pipelines and assets reference them by name.

Connection sets (for AI agents) ​

A connection set is a named bundle of connections used by an AI agent. Connection sets are kept separate from the connections your pipelines use, which lets you:

  • Restrict agents to only the data they need.
  • Give agents read-only access where pipelines have read/write.
  • Apply granular, agent-specific permissions without touching pipeline credentials.

To create a connection set, open AI → Connection Sets (in the AI Agent Settings sidebar), click New connection set, name it, and pick the data platform connections it should include. Then attach it to an agent from the agent configuration page.

Next ​