Consistent UGC AD Video Agent v1 in n8n

Search for a command to run...

No comments yet. Be the first to comment.
A fully automated voice-based dental receptionist that answers calls, books appointments, handles complaints, routes emergencies, and transfers to liv

This automated system turns one keyword into live SERP data, keyword ideas, and keyword suggestions—all neatly organised inside Google Sheets.

If you’re tired of spending hours every day searching for prospects, writing comments, and manually sending LinkedIn invites, you’re not alone.LinkedIn growth requires consistency — and consistency takes time. So I built a fully automated LinkedIn en...

Consistently posting valuable content on Twitter/X is one of the fastest ways to grow your audience.But coming up with daily ideas—especially in YOUR NICHE—takes time, research, and energy. To solve this problem, I built a fully automated Twitter/X i...


Download Workflow JSON: https://n8dex.com/2fnCXaER
Copy Google Sheet: https://docs.google.com/spreadsheets/d/1PVFbX_DkGf8l-bFCWXkTi_oDGgtSjPkPbRoI_Ryi6Pw/copy
Video Output:
If you want to go deeper, learn faster, and be part of a motivated community, you can join my skool:
👉https://link.thesohailjafri.me/join-skool
Or watch all my free automations on YouTube - AI With Sohail
For paid work contact: https://link.thesohailjafri.me/contact-ai
Reads Pending tasks from Google Sheets
Generates a merged UGC image using:
Product photo
Person photo
Text instructions
Generates Start Frame + End Frame prompts using NanoBanana format
Builds a VEO3-compatible script using a dedicated system prompt
Sends everything to KIE VEO to generate video
Polls until video is ready
Uploads images & collects final URLs
Updates success/failure back to Sheets
A schedule checks for rows where Status = Pending in the Google Sheet.
Pulls data from the row:
Product name, description, photo
Person name, description, photo
Target audience
Video instructions
Aspect ratio
The workflow formats all fields into a single combined string for model prompts.
Using OpenRouter (Gemini 2.5 Flash Image):
Creates a merged image containing the product + person
Uploads the merged image and stores URL
Using the NanoBanana 2-frame prompt generator:
Creates Start Frame + End Frame descriptions
Ensures character identity and product accuracy
Produces negative prompts for safety
The agent creates a clean, influencer-style selfie script:
Safety rules enforced (non-erotic, modest clothing, natural look)
Uses combined metadata + merged image
Returns a VEO-ready prompt + orientation
Sends the prompt + images to KIE VEO:
Model: veo3_fast
Mode: FIRST_AND_LAST_FRAMES_2_VIDEO
The workflow waits and checks until:
Video is ready → continue
Error code ≥ 400 → fail the task
On success → mark Completed + attach video URL
On failure → mark Failed
Add a new task to the Google Sheet row.
Fill in ALL required fields.
Set Status → Pending.
Wait for the scheduled run (or run manually in n8n).
Get your UGC video with consistent character + product branding.
If you want to go deeper, learn faster, and be part of a motivated community, you can join my skool:
👉https://link.thesohailjafri.me/join-skool
Or watch all my free automations on YouTube - AI With Sohail
For paid work contact: https://link.thesohailjafri.me/contact-ai