AIHub Expose Project Video

AIHub Expose Project Video reads a video file stored in the AIHub project and hands it to your workflow as a VIDEO object plus the raw LOCAL_FILE path. But read the docstring the way its author wrote it: "this util does not have the capacity to process the video file itself into something usable, it merely exposes the file name to be used by other nodes that can handle video files." That honesty is the whole personality of this node - it's a file gateway, not a video decoder.

Mechanically it's the project-file pattern: the workflow needs a project (via AIHubWorkflowController with a project_type), the client tells the server which file, the server injects the path, and the node wraps the file into ComfyUI's VIDEO representation by default. The README adds a sharper version of the same warning - "the expose project video function cannot load a video file and it must be done by another node" - which is why the escape hatch exists.

What you set

• file_name - the video file as stored, extension included (video.mp4 default).
• batch_index - pick one file from a stored batch by index; single integer, negatives count from the end.
• do_not_process - the important one. Check it and the node leaves the video alone entirely: VIDEO comes back null and you get LOCAL_FILE, which you feed to whatever node actually handles video. Set this when the default VideoFromFile wrapping isn't what you want, or when you need to process the file path yourself.
• optional - true means a missing file yields nulls instead of aborting.
• id - protocol field id.

Outputs: VIDEO, LOCAL_FILE.

The pragmatic reading

For most real workflows, do_not_process should be on. This node's value is that it gets a path to a project-stored video into your graph - the actual decoding and frame work belongs to a video-capable node downstream. Defaulting to processing is convenient for simple cases, but if you ever find yourself getting a VIDEO you can't do anything useful with, the flag is there precisely to get out of the way and hand you the file path instead.

Install & gotchas

Standard pack install: ComfyUI Manager → search ComfyUI-aihub-workflow-exposer, or:

cd ComfyUI/custom_nodes
git clone https://github.com/otavanopisto/ComfyUI-aihub-workflow-exposer

then restart. No requirements.txt, no model downloads.

Two gotchas. First, the no-project default trap - no project context, no file, and with optional off you get a hard "Video file not found" error. Second, don't expect this node to make frames: it doesn't, and no amount of wiring here replaces an actual video-processing node. If your workflow needs frames as images, that's AIHubExposeFrame (live from an editor) or a dedicated video loader (from files) - this node's job ends at getting the file into the graph.