Node LibraryFlow control

Run flow

Overview

Starts every node connected after it, so one click or one schedule runs the whole chain. It does no work of its own.

Use it as the entry point of a flow, especially when several independent branches should start together.

For example, use Run flow to launch multiple import branches in parallel, where each Import images node writes images into its own Dataset.

Run flow Import images Import images Import images Dataset Dataset Dataset

How it works

1

Completes immediately, without doing any work of its own.

2

Starts every node connected downstream, so parallel branches begin together.

FAQ

Select the node and press Run node in the sidebar. A schedule set on this node starts it the same way, except while flow scheduling is paused.

No. It starts only the nodes connected after it in the same flow, and each of those starts the nodes after it when it completes.

A node that still has a task running from an earlier run is skipped. So is a node with Require configuration turned on whose required fields are empty. Nothing connected after a skipped node runs either.

No. It has no inputs or outputs, and only starts the nodes connected after it.

Run flow fans execution out into parallel branches. Wait does the opposite: it joins branches back and continues only after all of them finish.

JSON config

Machine-readable node interface for automation and advanced usage.

{
  "name": "Run flow",
  "description": "Start the chain of nodes connected to this node. Performs no work itself.",
  "category": "Flow control",
  "namespace": null,
  "templateKey": "flow_control/run_flow",
  "version": "v1",
  "inputs": [],
  "outputs": [],
  "runtime": {
    "type": "headless",
    "requiresGpu": false,
    "dockerImage": null
  },
  "automation": {
    "manual_run": true
  }
}