Part 2: Deploy to the cloud

Both pipelines work locally, but you can't share local dashboards with your team. The dltHub Platform lets you deploy pipelines and dashboards to the cloud, schedule them, and share them with colleagues.

Log in

Connect your local workspace to the dltHub Platform:

uv run dlthub login              # device-code OAuth in the browser
uv run dlthub workspace connect  # pick or create a workspace

After connecting, open the platform UI:

uv run dlthub show

Every new account has a playground workspace. Your local workspace connects to it automatically, so anything you run locally syncs to the platform.

Deploy the pipeline

Tell the agent to deploy the REST API pipeline:

deploy this on the dlthub platform, use duckdb as destination

The agent installs the dlthub-platform toolkit. It goes through a five-step checklist before deploying, then registers the pipeline in __deployment__.py and deploys it.

You can also do it manually:

uv run dlthub deploy   # ship the current project as a new version
uv run dlthub run      # run the pipeline on the cloud

Repeat this deploy-and-run cycle after every code change so the cloud always reflects your latest version.

Ephemeral storage

When you deploy with DuckDB as the destination, the data goes to ephemeral storage. The platform runs your pipeline in a container, and when the job finishes, the local files are removed. The data doesn't persist across runs.

Switch to the Playground destination

To persist data, switch from duckdb to the playground destination, which is a managed S3 lake that keeps data across runs.

In rest_api_pipeline.py, change the destination:

# was:
#   destination="duckdb"
# now:
destination="playground"

The playground destination requires the deltalake package, so after changing the destination, redeploy and run it.

Then run the pipeline again:

uv run dlthub deploy
uv run dlthub run

If the run fails because deltalake is missing, the deploy step adds the dependency to pyproject.toml automatically. Redeploy and run again.

← REST API pipeline | Dashboard and scheduling →