Alibaba's HappyHorse 1.0 just dropped into ComfyUI, it's mainly for those e-commerce and short video final renders.

The Forge / 工坊
1

ComfyUI just dropped support for Alibaba’s HappyHorse 1.0, and it’s marketed as cinematic video generation. From the intro, it’s clearly aimed at practical use cases like ads, e-commerce visuals, and short videos.

They give you three generation paths: text-to-video, image-to-video, and subject-to-video (S2V). The selling point is that cinematic look—big aperture, shallow depth of field, that whole vibe. Max output is 15 seconds at 1080p, and they emphasize consistency across multiple camera cuts. There’s also V2V and SV2V editing flows, where you can swap or insert a subject while keeping the original motion and composition.

For those of us making detail-page videos, the image-to-video plus shallow depth of field combo is pretty spot-on—static product shots just come to life. But the real quality check will have to wait until we actually test it with real products. Just search “HappyHorse” in the template library and you’ll find a few workflows: image-to-video, text-to-video, video editing, and reference-based video generation.

2

Image-to-video is an absolute must-have for e-commerce.

3

That shallow depth of field vibe is exactly what e-commerce clients are after.

4

Waiting for real-world tests, I’m worried the product details will start morphing as it generates.

5

If S2V can actually keep the product consistent when generating videos, that’d be a game changer.

6

Here are a few possible translations depending on the tone of the original post:

Option 1 (Casual & Chatty):

mark

Option 2 (A bit more playful):

Marking my spot.

Option 3 (Short & Sweet, like a bookmark):

Bookmark.

Since the original is just “mark” (as in bookmarking a post), the most natural and common way a forum member would say it in English is simply “mark” or “bookmark”. I’ll go with the most direct and colloquial one:

mark

7

15 seconds at 1080p is more than enough for short videos.