n8n UGC Video Ads Workflow: GPT-4 Vision, Gemini & Sora 2

Turn a product photo into ready-to-generate UGC video scripts—then produce portrait ads with Sora 2

This n8n UGC Video Ads Workflow automatically takes a Product Image URL and Product Name from an n8n Form, uses GPT‑4o Vision to define an ideal UGC creator persona, has Gemini 2.5 Pro write multiple 12‑second UGC scripts, and then generates portrait videos for each script via the Prototipal Sora 2 gateway.

What this workflow does

• Receives inputs from an n8n Form Trigger: Product Image URL + Product Name.
• Downloads and encodes the image (converts it to base64) so it can be used as vision input.
• Analyzes the product image with OpenAI (GPT‑4o with image input) to generate a detailed UGC creator persona profile.
• Generates multiple scripts using Google Gemini 2.5 Pro, using the persona + product name + product image to produce three 12‑second UGC video scripts.
• Splits and iterates over the scripts, handling each one sequentially.
• Creates Sora 2 portrait video jobs through the Prototipal API, polls status every 15 seconds until completion, then downloads the resulting video.

Use cases

• Generate UGC ad creative for a SaaS or e-commerce product using only a photo and product name.
• Speed up testing of multiple 12-second hook angles for paid social campaigns.
• For automation engineers: streamline an AI pipeline from form submission → scripts → Sora 2 video outputs.

Technical details

• n8n nodes: if, set, wait, split out, sticky note, form trigger
• OpenAI Vision: GPT‑4o image input for persona generation
• Google Gemini: Gemini 2.5 Pro for script generation
• Prototipal Sora 2 gateway: job create → status polling (every 15s) → download