This @Scenario_gg Workflow takes any image (character, object, or scene) and automatically extracts all individual elements (typically 10–15) in one single click. It use an LLM node to identify & list each element, then SAM 3.1 from @Meta to mask them, @OpenAI GPT Image 2 + Pixa background removal to reconstruct and isolate. From there, you can group everything into a grid or keep iterating (rearranging elements, more image editing, 3D conversion, or video gen...)

@emmanuel_2m @Scenario_gg @Meta @OpenAI Nice. Been trying to do this for a while and results are hit and miss. Will give it a go and then I guess next step is adding overdraw so can wire back into a spine animation