Sogni AI icon

Sogni AI

Generate AI images, videos, and LLM responses using Sogni AI Supernet

Actions18

Creative Workflow → Start

AI-generated

Summary

Start a creative workflow run by specifying either a saved workflow template ID with optional input parameters, or a fully defined inline workflow JSON object.

Inputs

  • Start Mode (required) — Choose between running a saved template ('template') or an inline one-shot plan ('inline').
  • Template ID (required) — When using 'Saved Template' mode, the ID of the workflow template to run.
  • Inputs JSON — Optional JSON object to provide input values to the saved template's declared inputs.
  • Inline Workflow JSON (required) — When using 'Inline Plan' mode, provide the full workflow definition as a JSON string containing properties like title and an array of steps with their tool name and arguments.
  • Wait Until Terminal — Boolean flag to poll and wait until the workflow reaches a terminal status (completed, failed, etc.) before returning. Defaults to false for immediate return of the queued run record.
  • Start Additional Fields — Optional collection of advanced parameters including token type (Spark or SOGNI), application source tag, idempotency key for safe retries, caps on capacity units, cost confirmation flag, media references JSON array for workflow media usage, and polling interval/timeout settings.

Output shape

a single record representing the started creative workflow run

If 'Wait Until Terminal' is false, returns immediately with the queued run record. If true, polls the run status at intervals until reaching a terminal state or timing out. Optional fields and parameters influence run execution details, cost controls, and media usage.

Links

Discussion