Upload Video

The node that gets an existing clip onto Animon's server

Animon Upscale Video doesn't accept a file or a tensor - it wants an ANIMON_VIDEO_ID, a handle to a video already uploaded to Animon. Animon Upload Video is how you create that handle from a file you already own. It's the video twin of LoadImage, but with a cloud upload where the file picker used to be.

Think of the flow: you've got an old anime clip, a generated render, or a piece of footage you want cleaned up to 1080P. You drop it in ComfyUI's input folder, pick it from this node's dropdown, and the VIDEO_ID that comes out feeds straight into the upscale node.

How it works

Like its image sibling, the node lists video files from folder_paths.get_input_directory() (ComfyUI's input folder), filtered by content type, and shows them in the video combo. On execution it opens the file and calls Animon's files.create with purpose="video_generation", returning the server's file ID as VIDEO_ID. That's the entire job: local file in, cloud handle out.

Inputs are just animon_key and the video picker. If your clip isn't in the dropdown, move it into ComfyUI/input/ and refresh the node.

Installing and the constraints that matter

Standard pack install: ComfyUI Manager → search ComfyUI-Animon → restart, Key node wired up. Requirements are only openai and decord.

The constraint to actually respect is the upscale node's, because that's usually why you're here: the input must be smaller than 1080P and no more than 360 frames. Uploading a 4K file or a ten-minute clip just means the upscale call will fail server-side - trim and resize first, then upload. And if the video you want to upscale was just generated by Animon itself, skip this node entirely: Upload Video (from Bytes) takes the generation node's VIDEO_BYTES output directly, avoiding a re-encode and a second trip through your connection. This file-based node is the right tool when the clip started life outside the Animon pack.