Skip to main content
Every pipeline has exactly one Trigger. It defines when the pipeline runs.

​
Manual

The default. Pipeline runs only when you click Run in the canvas, or from the CLI:
Good for ad-hoc backfills or testing. Not recommended for anything you want kept in sync.

​
Scheduled

Runs on a cron. Choose:
  • Cron expression — 5-field (0 * * * * = hourly). Common presets in the picker.
  • Timezone — defaults to workspace timezone.
  • Skip if previous still running — on by default. Prevents backlogs.
The next scheduled run is shown on the Trigger block so you always know.

​
Watch

Watches an external location and runs whenever it changes:
  • Watch a folder — Google Drive folder, S3 prefix, Notion database. Change detection is content-hash based, so you don’t re-process unchanged items.
  • Watch a channel — Slack channel, Discord channel. Runs per new message batch.
The polling interval defaults to 5 minutes. Real-time push subscriptions are used where the connector supports them (Google Drive, Slack).

​
Webhook

Turns your pipeline into an HTTP endpoint. Any external system that can POST JSON can trigger a run.
  • URL: https://api.platform.nora.my/pipelines/{slug}/trigger
  • Signing secret — HMAC-SHA256 in X-Nora-Signature.
  • Payload — passed to the Source stage as the run input.
Useful when you already have a webhook available from your CMS, ticketing system, or CI.

​
Passing config to a run

All Trigger types accept an optional run-time payload. The payload becomes an input variable available in every stage of the pipeline as {run.payload.KEY}. Use it for:
  • Manual runs targeting a single record: { "record_id": "abc123" }
  • Scheduled runs that need date parameters: { "since": "2026-07-01" }
  • Webhook payloads with source-of-change info.

​
Pausing a pipeline

Toggle Paused on the pipeline. Scheduled and Watch triggers stop firing. In-flight runs finish. Publishing a paused pipeline unpublishes it from production Agents until you unpause.