Skip to content
FlixMLFlixML Studioopen-source generative AI image and video tool

SAM 3 Video Mask

Cuts one subject out of a video as a mask: the subject white, everything else black, same frame count and frame rate as the clip. SAM 3 finds what the prompt names and tracks it through every frame. For motion-capture nodes that take a person mask video alongside the clip. Uses ComfyUI's core SAM 3 nodes, the same tracking step as its SAM 3D Body template. SAM License: https://github.com/facebookresearch/sam3/blob/main/LICENSE

Workflow id: sam3_video_mask · Video → Video

Calling it

POST /api/video/generate
{
  "workflow": "sam3_video_mask",
  "prompt": "...",
  "video": "..."
}

Params

Every input this workflow takes. Your install is the source of truth:GET /api/workflows/sam3_video_mask returns these as they are on your machine.

  • prompt · str · required
    What SAM 3 tracks, in a word or two. With more than one person in frame, name the one you want ("man", "mannequin", "woman in red").
  • video · str · required
    The clip to mask, as a ComfyUI input filename.
  • max_objects · int · default 1
    Most matches SAM 3 tracks. 1 keeps only the best match for the prompt.
  • sam3_model · str · default sam3.1_multiplex_fp16.safetensors
    models/checkpoints. https://huggingface.co/Comfy-Org/sam3.1
  • filename_prefix · str · default videos/sam3_video_mask