UGC workflow (with VEO 3.1)

Video is being edited. Coming soon...

This system turns product images into UGC-style video ads automatically using Veo 3.1 (via Fal AI) and AI agents in n8n.

The goal: replace expensive, slow UGC production (creators, shoots, editing, revisions) with an automated pipeline that creates authentic, “creator-style” ads from simple inputs:

  • product name
  • one image (text-to-video)
  • or two images (frame-to-video start & end)
  • basic video description

The output: ready-to-use UGC-style video ads that look and feel like real creator content you can plug directly into your paid campaigns.

The system supports two main modes:

  1. Text to Video –

    • Upload one product image
    • Describe the type of UGC video you want
    • The workflow analyzes the product look, generates a UGC-style image, then animates it with Veo 3.1 into a testimonial-style video.

  2. Frame to Video –

    • Upload two images: a start frame and an end frame
    • Describe the desired motion
    • The workflow analyzes both frames, designs motion between them, and uses Veo 3.1 to generate a smooth transition video.

Everything is orchestrated in n8n using:

  • Form Trigger → user interface
  • Google Drive → storing images & videos
  • OpenAI (GPT-4o) → product image analysis
  • Claude Sonnet 4 → image prompt + video motion prompt
  • Fal AI (Veo 3.1 endpoints) → image & video generation
  • Wait / status polling → reliable async generation

No manual editing, no timeline work, no creator coordination.

What You Need to Customize

Before using this workflow in your own stack, you’ll need to adjust a few key parts:

1. API Keys & Connections

Update credentials in n8n:

  • Fal AI API key (for image + video generation with Veo 3.1)
  • OpenAI API key (for GPT-4o vision analysis)
  • Anthropic API key (for Claude Sonnet 4 agents)
  • Google Drive credentials (for file upload & retrieval)

Make sure each node is pointing to the correct credential.

2. Form Setup & Options

The Form Trigger should match your use case:

  • Fields:
    • Product Name
    • Video Details (what you want the UGC ad to show/feel like)
    • Format (dropdown: "text to video" / "frame to video")
    • Single image upload (for text-to-video)
    • Two image uploads (for frame-to-video)

If you rename any options (e.g., “Text → Video”, “Frames → Video”), update the Switch node conditions accordingly.

3. Google Drive Folders

In all Google Drive nodes:

  • Set your folder IDs for:
    • Uploaded product images
    • Uploaded frame images
    • Final video output

If you change folder structure later, remember to update the node settings.

4. Prompts & Brand Style

You’ll likely want to adapt the system prompts to your niche/clients:

  • Image Prompt Writer Agent (Claude):

    • Describe what your UGC should look like (casual vs premium, selfie-style vs B-roll, etc.)
    • Define target audience, tone, environments, and typical scenarios.

  • Video Prompt Writer Agent (Claude):

    • Tune motion style: more dynamic vs subtle, handheld vs smooth, close-ups vs mid-shots, etc.
    • Add constraints around duration, energy, and pacing to match your ads.

This is where you differentiate your system from a generic template.

5. Fal AI / Veo 3.1 Calls

In the HTTP Request nodes to Fal AI:

Double-check:

Endpoint URLs (image vs video)

Model name (Veo 3.1)

Request body (prompt, references, aspect ratio, duration, quality)

Make sure:

  • Text-to-video path sends:

    • The generated image URL as the starting frame
    • The video prompt JSON from the Video Prompt Agent

  • Frame-to-video path sends:

    • Both start & end frame URLs
    • The transition/motion JSON from the Video Prompt Agent

6. Wait & Status Polling

Adjust Wait durations and retry logic:

  • For images: ~20–30 seconds is usually enough
  • For video: ~30–60 seconds per retry

If your jobs frequently take longer, increase the Wait duration or max retries to avoid premature failures.

Once these elements are customized, the system runs end-to-end:

  • Fill out the form →
  • Routing based on format →
  • Analyze product and/or frames →
  • Generate UGC-style image (if text-to-video) →
  • Generate motion prompt →
  • Generate Veo 3.1 video →
  • Store final file → ready-to-use UGC ad.

You go from product image → UGC-style video ad in a few minutes, for pennies, with no creators involved.

← Previous
Next →