Okay, so this isn’t precisely what I thought it was at first (video inpainting), but rather an creation->inpainting->animation flow. Still, the results look impressive:

How it works:

→ Generate an image in Higgsfield Soul
→ Inpaint directly with a mask and a prompt
→ Combine with Camera moves, VFX, and Avatars to turn static edits into living, speaking visuals pic.twitter.com/ENHqdA3WHm

— Higgsfield AI (@higgsfield_ai) July 3, 2025