Load Video
The entry point that hands you a string, not a video
- video
Load Video is the front door of the ComfyUI Core Video Nodes pack (Immac/ComfyUI-CoreVideoMocks) - and I have to get the bad news out of the way first. This is the pack's mock of what a native ComfyUI video loader would look like. It's an RFC proposal, not a real loader. The GitHub repo describes itself as "A set of mock nodes for RFC purposes," and this node lives up to that billing perfectly.
In a real pipeline this is the node that reads a video file off disk and gives the rest of the graph something to work with. Here's what actually happens: you pick a file, the node returns a single string like "Mock Loaded Video: myclip.webm", and that string is the whole output. No frames are read, no codec is probed, no pixels exist. The video is a label, not a video.
The one input
The only input is video, a dropdown that lists every file in your ComfyUI input folder (it also shows an upload button, thanks to image_upload). So you do get a genuinely useful behavior: a real, populated list of your input files. The node even validates that the file still exists before it runs. It's just the "loading" part that's fake.
How to install it
Same as every node in this pack:
cd ComfyUI/custom_nodes
git clone https://github.com/Immac/ComfyUI-CoreVideoMocks
Restart ComfyUI. No pip requirements, no models, no ffmpeg - it's pure Python on top of ComfyUI's own folder_paths. ComfyUI Manager also finds it if you search "ComfyUI Core Video Nodes."
Why you'd bother
Honestly? Mostly for learning. This node's job in the pack is to show where a "load" step fits in a proposed core video workflow - load, split into streams, decode, edit, encode, save. If you're sketching out how ComfyUI's video API could be organized, it's a useful shape to study. If you actually want to load a video and turn it into something you can edit, reach for VideoHelperSuite (VHS) instead - that's the pack people actually run for real video work, and unlike this one it hands you actual frames.
One honest warning if you do wire it up: every node in this pack feeds placeholder strings downstream, so a Load Video → Decode Video Stream chain will "run" and produce nothing you can use. Treat the pack as a spec document with a UI, not a pipeline.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| video | COMBO | 1 options: example.png |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| video | STRING | — |