Bruin CLI

Appearance

Getting Started ​

Bruin Cloud has two tracks. Pick the one that matches what you want to do first — you can set up the other later from the same workspace.

AI

AI Data Analyst

Connect a warehouse, ask questions in plain English, get charts and reports back. Push answers into Slack, Teams, or WhatsApp.

  • Best for analysts and PMs
  • Connect once, ask anything
  • No code required
Start with AI Data Analyst →
ETL

ETL/ELT Pipeline

Define SQL and Python transforms in Git, then schedule them. Bruin parses, validates, and runs your pipelines automatically.

  • Best for data engineers
  • Git-based, code-first workflow
  • Scheduled SQL + Python transforms
Start with ETL/ELT Pipeline →

Before you start ​

Sign up at cloud.getbruin.com/register with email and password, or sign in with Google. Right after sign-up, Bruin Cloud asks which track you want — pick AI Data Analyst or ETL/ELT Pipeline.

If you'd rather start in code, the open-source Bruin CLI defines the same pipelines, assets, and connections Bruin Cloud runs. Start with the Quickstart, then connect your repo to Cloud.

TIP

You can switch tracks at any time from Getting Started on the home page. The toggle is labelled AI Analyst / Data Engineer.

AI Data Analyst track ​

Connect a warehouse, ask the agent questions, then push answers into your team's chat tools.

1. Connect your data. Add a connection to a warehouse the agent should read from — BigQuery, Postgres, MySQL, SQL Server, Snowflake, Databricks, or Redshift. Create with validation so Bruin can confirm the credentials work.

TIP

No warehouse credentials yourself? Invite a teammate from your data team and have them set the connection up. The agent then becomes available to the whole workspace.

2. Ask your first question. Open AI → Chats, pick the default agent, and try:

What data do you have access to?

The agent inspects the schema and tells you what's available. From there, ask the questions that matter — revenue, retention, top SKUs, whatever your team needs. See Chat with Agents.

3. Connect a chat platform. Open AI → Agents, pick the agent, and wire it to where your team already works:

  • Slack — OAuth install, then a Channel ID per agent
  • Microsoft Teams — connect BRN-XXXX in a channel, group chat, or 1:1
  • Google Chat — connect BRN-XXXX in a DM or space
  • Discord — bot install, then /bruin in any wired channel
  • WhatsApp — message Bruin's number with connect BRN-XXXX
  • Telegram — DM @BruinDataBot with connect BRN-XXXX

4. Build a dashboard. Open AI → Dashboards and describe what you want. The agent assembles charts, tables, and metrics from your prompt; iterate until it looks right, then click Publish. See Dashboards.

5. Schedule a report. Have the agent run on a cadence and post results into your chat tool. Daily revenue summaries, weekly retention alerts, threshold-based notifications. See Scheduled Agents.

6. (Optional) Add a context layer. Connect a Git repo with your dbt or Bruin semantic layer so the agent learns your team's vocabulary and metric definitions. Or describe tables manually for the metrics that matter most. Both options are in Team Settings → Projects.

ETL/ELT Pipeline track ​

Wire a Git repo, configure the connections your pipelines need, enable them, then monitor.

1. Create a project. A project in Bruin Cloud maps one-to-one with a Git repository. Creating one syncs the pipelines in that repo and gives you a place to manage their connections.

TIP

Use the Bruin GitHub App instead of a personal access token where possible — fine-grained access, no expiring tokens, repo-level installs. See Projects.

2. Add connections. While the project syncs, open Connections and add the data sources, destinations, and secrets your pipelines reference. These are the same connections you'd define locally in .bruin.yml, but encrypted in Bruin Cloud's HashiCorp Vault instead of your repo.

See Connections for connection types, naming rules, and the validation flow.

3. Enable a pipeline. Pipelines synced from a repo start disabled. Open Catalog → Pipelines, pick one, and click Enable selected pipelines. If any referenced connections are missing, Bruin Cloud lists them and lets you add each one inline.

INFO

The first run triggers automatically when you enable a new pipeline. You don't need to click New run yourself.

4. Monitor and operate. Once a pipeline is active, the pipeline page is your operations console:

  • Runs panel for status and history. See Runs.
  • Assets for every asset with its type, owner, schedule, last run. See Assets.
  • Backfills for historical reprocessing. See Backfills.
  • Lineage for how assets connect. See Catalog → Lineage.
  • New run for ad-hoc runs, full refreshes, and backfills.

5. Set up notifications. Configure pipeline-level alerts on success or failure in your pipeline.yml. Bruin Cloud supports Slack, Microsoft Teams, Discord, and generic webhooks. See Notifications.

6. (Optional) Cross-pipeline dependencies. If pipelines in different repos depend on each other's assets, wire them up with URIs instead of duplicating definitions. See Cross-pipeline dependencies.

After the first run ​

These apply regardless of which track you started with:

  • Team Settings — members, projects, billing, audit logs
  • API Tokens — programmatic access (CI, MCP, external monitoring)
  • Insights — cost explorer, pipeline health, risk report, usage
  • Cloud MCP — talk to Bruin Cloud from Cursor, Claude Code, or Codex
  • FAQ — common questions and patterns that aren't real features