Modelverse HappyHorse T2V

HappyHorse 1.0 is the budget lane of this pack's video lineup. The name is not a promise about output quality - it's the model's actual name - and honestly it sits in the pack as the "don't burn Sora credits on a first draft" option. Prompt in, mp4 out, through UCloud's Modelverse API, no local GPU involved. When you're iterating on an idea and just need to see the motion work before committing to a pricier frontier model, this is the node.

What it is

A straight text-to-video generator (happyhorse-1.0-t2v under the hood). One prompt, a resolution tier, an aspect ratio, a duration, and you get a video back. No frames to supply, no reference material, no faff. Think of it as the quick-sketch mode of the video pipeline: it won't win film festivals, but it will tell you if your idea has legs in seconds rather than quarter-hours.

The inputs that matter

• client - required, from the Modelverse Client / Secret Client node.
• prompt - the video description. This is everything; write it with motion in mind.
• resolution - 480P, 720P (default), or 1080P. The 1080P tier costs more per run.
• ratio - 16:9, 9:16, 1:1, 3:4, or 4:3.
• duration - 1–15 seconds, default 5. Longer = slower and pricier.
• seed - 0 for random, or fix it to reproduce a take.
• watermark - on by default? No: off. Flip it if you want the branding.

Outputs are the pack-standard url + task_id strings. The video becomes visible when you run url through the Modelverse Preview Video node, which downloads it to ComfyUI/output and hands you a VIDEO tensor for preview and further processing.

Installing and keying up

Every node in this pack lives in the one plugin. Install via ComfyUI Manager (search "ComfyUl-UCloud" → Install → restart) or:

cd ComfyUI/custom_nodes
git clone https://github.com/ucloud/comfyui-plugin

restart ComfyUI. Dependencies are just requests, pillow, numpy, openai, GitPython - Manager installs them; there are no model weights to download, because the model never touches your machine. You need a ModelVerse API key from the UCloud console: either type a secret name into the Modelverse Secret Client node (key stored in secrets.json, out of the workflow) or add MODELVERSE_API_KEY=... to the pack's config.ini.

Common issues

• Videos look "meh" at 480P - that's the model + tier doing its job. Use 720P as your real baseline and treat 480P as pure draft mode.
• Nothing to preview - remember the video node outputs a URL string, not pixels. It has to pass through Modelverse Preview Video.
• "API key is not set" - empty client node, every time. Check the client, not the video settings.
• Long queue for a short clip - even quick models poll every 5 seconds and can sit in the queue while the cloud spins up capacity. A 5-second draft can still take a couple of minutes wall-clock.

It's the pragmatic pick: iterate in HappyHorse, and only when the take is right do you send it up the ladder to Sora or Kling. The people who get value out of this pack are the ones who treat the cheap nodes as the sketchpad.