Quickstart — Nextflow CLI

One command launches the pipeline and tails its progress right there in your terminal:

ccloud nextflow run nf-core/demo --outdir s3://my-bucket/pipeline-runs/demo-1/out -profile test

Under the hood this is the same launch API as Quickstart — API: each run gets its own ephemeral head, created for that one run and torn down automatically when it closes. The CLI just writes the JSON for you and keeps watching.

When to use this

• You live in a terminal and want the fastest path from “idea” to “running pipeline” — no head to create, nothing to SSH into.
• You want launch + live progress in one command, with the full DAG a browser tab away.
• You’re scripting casually (a Makefile, a cron job) but don’t want to hand-roll API calls — for serious automation or reselling, use the API quickstart.

1. Install the CLI and set your key

curl -fsSL https://console.carolinacloud.io/static/cli/install.sh | bash
export CCLOUD_API_KEY="your-api-key-here"

Generate the key from the dashboard under Settings > API Key (see the CLI reference for details).

2. Point it at a bucket you own

--outdir is required and has to be an s3:// path in a Carolina Cloud bucket you own — your auto-provisioned pipeline bucket, a personal bucket, or an org bucket (org buckets: org admins only). See Storage & buckets. Don’t have one handy? ccloud buckets list shows every bucket you can write to.

3. Launch

ccloud nextflow run nf-core/rnaseq \
  --outdir s3://my-bucket/runs/rnaseq-1/out \
  --input samplesheet.csv \
  -profile test -resume

Everything after the pipeline name is passed through, with Nextflow’s own convention:

• Single-dash flags (-profile test, -resume) become Nextflow engine flags.
• Double-dash flags (--input samplesheet.csv) become pipeline params.

A few flags belong to the CLI itself:

Flag	What it does
--outdir	Required. s3:// destination — see step 2
--config <file>	A local nextflow.config, read and shipped to the head (layered on top of the head’s own executor config)
--pipeline-type	ref (default — an nf-core name or git URL) or script (submit a main.nf directly)
--json	Print the raw launch response and exit instead of watching the run

Local files just work

Any flag value that’s a local file — a samplesheet, a params file — is detected and staged onto the head automatically (up to 5 MiB combined). The file’s contents must still point at object storage (s3:// paths); the CLI never uploads what the file references. -c/-config/-plugins are reserved by the server — use --config for config, and the nf-ccloud executor plugin is loaded for you.

4. Watch it run

The launch prints the run id and console link, then settles into a live status line (the same data the console DAG polls):

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  🧬  Pipeline launched: nf-core/rnaseq
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

  Run     hopeful_darwin
  Run ID  45d91a6b-64b9-4711-b58b-727ad0d98d90
  🔗 https://console.carolinacloud.io/nextflow/45d91a6b-64b9-4711-b58b-727ad0d98d90/

Ctrl+C does not kill the run. It pauses and asks whether you want to cancel the run server-side or just detach and leave it running — so closing your laptop mid-run is safe either way. A detached run is still visible in the console, and in GET /api/pipeline-run/ if you’d rather poll.

5. If it fails

The status line reports the failing task and Nextflow’s error. For the full story, the debug log endpoint returns an LLM-ready bug report — see the API quickstart’s debug-log step; it works for CLI-launched runs too, same run id.

Next

• Watching a run — the console DAG viewer, task drill-down, and work-dir browser for this same run.
• Storage & buckets — bucket ownership, credentials, and the one-endpoint rule.
• CLI reference — everything else ccloud can do.