Skip to main content

Quickstart

Get a pipeline running on dltHub Runtime in under 10 minutes.

Prerequisites​

  • Python 3.10+ and uv installed
  • A dltHub Runtime invite code (add it to .dlt/secrets.toml)
  • A MotherDuck account (or another supported destination)

1. Clone the starter pack​

git clone https://github.com/dlt-hub/runtime-starter-pack.git
cd runtime-starter-pack
uv sync

See the Starter Pack page for a tour of what's included.

2. Configure credentials​

Add your Runtime invite code and destination credentials to .dlt/secrets.toml:

[runtime]
invite_code = "your-invite-code"

[destination.motherduck.credentials]
database = "my_db"
password = "your-motherduck-token"

For profile-specific configuration (e.g., read-only credentials for notebooks), see the Manage Secrets guide.

3. Log in​

Authenticate with Runtime via GitHub OAuth:

dlt runtime login

This opens your browser for authentication and connects your local workspace to the remote one. See dlt runtime login for options.

4. Run your first pipeline​

Deploy your code and launch the fruitshop pipeline:

dlt runtime launch fruitshop_pipeline.py

This syncs your code and configuration to Runtime, creates a job, and starts a run. Add --follow to stream logs until completion:

dlt runtime launch fruitshop_pipeline.py --follow

See the Deploy a Pipeline guide for more detail.

5. View results​

Open the Runtime dashboard in your browser:

dlt runtime dashboard

You'll see your workspace overview with the pipeline run results, metrics, and logs.

6. Try an interactive notebook​

Deploy the fruitshop notebook as an interactive app:

dlt runtime serve fruitshop_notebook.py

This starts a Marimo notebook in the cloud. Once ready, it opens in your browser. See Serve a Notebook for more.

Next steps​