Sydney Sweeney Consistant Character | AI Music Video

Testing consistent character generation in an AI music video workflow — fully automated, with no manual editing.

AI agent workflow inside ComfyUI, where the agent generates images using an open-source image model, then uses LTX 2.6 for img-to-video generation.

For character consistency, I trained a Z Image LoRA of Sydney Sweeney.
It worked well for testing consistency, and with an LTX LoRA linked in, it could get even closer to the target character.

Big thanks to vrdevgirl for the workflow

I tested the workflow on a 4090, and the full render took almost 3 hours.
That gives a good idea of how powerful — and computationally intensive — these systems are right now.

If you’re in the music video industry, building with AI, or curious about where this is heading, comment ComfyUI and I’ll send you the GitHub workflow link.

Видео Sydney Sweeney Consistant Character | AI Music Video канала Govind Singh