n8n Workflow: UGC Video Request to Discord Approval

Turn UGC video requests into approved AI prompts—then generate the video automatically

This n8n workflow captures a UGC video request from an n8n Form, uses OpenRouter to extract product details and generate structured image/video prompts, sends a preview for Discord approval, and—once approved—generates the UGC video with Replicate (minimax/video-01). Every run is logged to an n8n Data Table.

What this workflow does

• Collect inputs from an n8n Form: a product image, a UGC prompt, and a selected video style.
• Analyze the product image with OpenRouter to perform product-detail extraction.
• Generate structured prompts using an OpenRouter chat model: clean UGC image/video prompts plus negative constraints, returned as parseable JSON fields.
• Create a prompt preview using the Fal API (fal-ai/nano-banana-2/edit), then post the preview to Discord for human review.
• Review decision loop: a second n8n Form captures the reviewer’s decision.
  • If approved: trigger video generation.
  • If feedback is provided: regenerate prompts using reviewer feedback.
  • If rejected: stop the job.
• Generate the video with Replicate (minimax/video-01), poll the prediction until it succeeds, and post the final video URL to Discord.
• Log results with job metadata in an n8n Data Table and send a completion notification to Discord.

Use cases

• Brand teams requesting UGC variants for new product pages, with a reviewer approval step via Discord.
• SaaS operators streamlining UGC production: form submission → AI prompt drafting → approval → Replicate video generation.
• Automation engineers building a controlled pipeline for AI-generated marketing assets with human-in-the-loop governance.

Technical details

• OpenRouter credentials for both product analysis and the OpenRouter chat model prompt generation.
• Fal API using an HTTP Header Auth credential for https://fal.run/fal-ai/nano-banana-2/edit.
• n8n nodes/flow: set, code, form, wait, switch, discord.