- Robin van Veen - AI Automation
- Posts
- How I automated viral ASMR videos with AI
How I automated viral ASMR videos with AI
Maybe you’ve seen them already:
Those oddly satisfying ASMR videos where a hydraulic press flattens something in perfect detail.
I just built an AI agent that makes these videos automatically — every single day — with zero editing required.
This is what’s possible now.
Using VO3, Google’s new text-to-video model, you can generate realistic, shareable videos on autopilot.
And it’s not just about making content for fun.
You can plug this into your workflows to grow an audience, feed social channels, or even experiment with monetization.
In my new AI agent build video, I show you step by step how I built this AI agent inside n8n:
✅ It picks a color and an object to feature in the video
✅ Generates a detailed text prompt describing the scene
✅ Sends the prompt to VO3 via an HTTP request
✅ Waits for the video to be created (takes about 5-10 minutes)
✅ Saves the final video URL in Google Sheets
✅ Keeps track of which combinations you’ve already used
✅ Automatically prepares your next video idea
I even walk you through how to:
→ Make sure every video starts right at the action to keep costs down
→ Use a prompt structure that looks hyper-realistic
The first time I ran it, it cost about $6 to create an 8-second video — but honestly, the result was way better than I expected.
Watch the video here:
Download the full JSON files here: