Distributed Video Placeholder

The video version of the pack's placeholder trick. No inputs; one video output that is a single black 1×1 frame at 1 fps, built entirely on the CPU. It exists for exactly the same structural reason as DistributedPlaceholder does for images: when the master's "master rendering enabled" setting is off, the master's copy of the workflow still has to execute to coordinate the job, and any downstream video node still expects a VIDEO to flow through it. This node stands in for the whole generation chain - instant, black, no GPU involvement - while real clips render on the workers and come back through a video collector.

There's nothing to configure, and that's the feature. The source builds a zero tensor, wraps it in the ComfyUI native VideoComponents structure via the same _get_video_api() helper the video collector uses, and hands it back. Pure plumbing.

Reach for it only in the specific situation where the master coordinates a video job without rendering anything itself. If your master renders too, or if you're running a hand-configured master/worker setup with rendering left on, you won't need it. And don't confuse it with a preview or a black-video filler - it's a structural shim, not an output. The Deadline-prefixed DeadlineDistributedVideoPlaceholder is the same node under the studio-friendly name, used when you want your Deadline submission workflows to read consistently.