Agentic nodes run a multi-step AI pipeline — reason about the best model (optionally researching), generate, then auto-QA the result — within a budget you set. Agentic Image and Agentic Video share this flow; you give a brief, references, and limits.
Where: flowGen → left rail → Add → Image → Agentic Image Gen, or Video → Agentic Video Gen — Add an agentic node
Key ideas
- Pipeline: reasoning → auto model-pick → execute → final QA (auto-scored), with optional web research.
- Controls: Max steps, retries/step, max budget (Flowies), “ask me”, and (video) duration + aspect ratio.
- Output: A versioned result with its QA score and Flowie cost.
Steps
- Add an Agentic Image or Video node.
- Give it a brief and references; set budget/steps.
- Run — it picks a model, generates, and QA-scores.
- Review the versioned result.
Pick the agent scope
- Agentic Image: one image with optional research, retry, QA, and budget.
- Agentic Video: one clip with typed references, retry, QA, duration/aspect, and budget.
- Agentic Long Video: coordinated shot planning and continuity.
- Agentic Long Video Edit: source-footage restructuring.
Limitations and important notes
- Auto-QA is a ranking signal. Review factual accuracy, rights, brand fit, identity consistency, and the full temporal result yourself.
Related guides
- Flowgen canvas: cards, node graph, typed inputs, and running a workflow
- Generation Providers & Models
- Generate and evaluate an image with Agentic Image
- Generate and evaluate one clip with Agentic Video
- Plan and generate a multi-shot video with Agentic Long Video
- Restructure footage with Agentic Long Video Edit