Automating NanaBanana images & VEO 3.1 Fast clips — but it keeps breaking. Help?
I need your help — I've been building automations to generate NanaBanana images and VEO 3.1 Fast video clips using Google Flow, but I'm running into a wall and I'm not sure what the best solution is.
Here's the problem:
My current automation was built with Claude Code and works by clicking through the Google Flow UI. The issue is it's completely brittle — if Google Flow updates its interface even slightly (a button moves, a new feature shows up), the whole thing breaks. There's no self-healing, no intelligence. It just fails silently.
What I've explored:
  • Google AI Studio API — I know I can use it to generate NanaBanana images, and that would be rock solid. But using it for VEO 3.1 Fast video generation is going to be extremely expensive at scale.
  • UI automation — Works until it doesn't. Not reliable long-term.
What I'm looking for:
Is there a solution that is:
  1. As reliable and fast as a direct API connection
  2. Smart enough to see the UI and adapt if something changes
  3. Cost-effective for video generation (VEO 3.1 Fast specifically)
Basically — something in between "dumb UI clicking" and "full API at full price."
Has anyone cracked this? Any tools, workarounds, or smarter approaches you're using? Drop them below 🙏
1
1 comment
Khalid Alanazi
3
Automating NanaBanana images & VEO 3.1 Fast clips — but it keeps breaking. Help?
AI Automation Society
skool.com/ai-automation-society
Learn to get paid for AI solutions, regardless of your background.
Leaderboard (30-day)
Powered by