A.I. can finally create REAL emotion on screen!!! Character consistency in VEO3 is still tricky.
Here’s how I did it…

I’m always happy to connect, so feel free to DM me—I’d love to chat!

So, honestly, it’s not really a hack, but right now it’s nearly impossible to get real character consistency with VEO3 alone. The workflow I use for all my films (spec work and client projects) goes like this:

1. I generate the key scenes with Google DeepMind VEO3 (text-to-video).

2. I grab screengrabs and modify them using Runway Gen-4 & Flux context.

3. That’s how I get character consistency and then create the close-ups with Kling AI 2.1 Master (img2video).

This way, I can maintain consistency within a scene, since I always generate the wide shots with VEO3 and the close-ups with Kling 2.1 Master.

Quick note on the cost:

For this project, I actually did some shots with VEO2, because VEO3 is just WAY too expensive. I used up all my beta tester credits on Freepik & Hedra (thanks for that!!!) and even paid for extra credits out of my own pocket. I’m sure all this will get a lot cheaper really soon, but as of now, you really have to think about which shots you want to do in VEO3 and which ones are better to make in Kling.

Overall, I’m super happy that I can finally make an ad that REALLY conveys emotion. What do you guys think? How much longer will we still be filming, and when will everything just be generated?

Produced by @max.schlett

Edited by @dylan_disengaged

Music Roza – I’ll Understand (via Artlist)

Tools I used:

Freepik , Hedra , Runway Gen4, Google DeepMind VEO3, VEO2, Kling AI 2.1 sync. Black Forest Labs Flux Kontext