Generate a flam
Write a prompt describing the motion you want, optionally connect image or video references, and click Run.Frames vs. reference inputs
Connected images and videos play one of two roles, set on the node:- Frames — first/last conditioning: the clip starts on your first image, and optionally ends on the second. Maximum of 2 inputs. Use it to bring a specific still to life.
- Reference — style and subject references: up to 9 images and videos condition the look without pinning the start frame. Use it to keep a character or product consistent while the flam invents its own motion.

