Alibaba video model

HappyHorse 1.1. Start from words, a frame, or references.

Alibaba's HappyHorse 1.1 turns a text prompt, a single starting image, or a set of reference images into video. Clips run 3 to 15 seconds in seven formats, from 21:9 ultrawide to tall 9:21.

Try HappyHorse 1.1->See all models

3 to 15 second clips
7 aspect ratios
First frame control
Reference images
Sound with the picture
Up to 1080p

Briefs arrive in different shapes. Some are a sentence, some are a product still that has to open the shot, and some are a handful of photos of a subject that must stay recognizable. HappyHorse 1.1 takes all three. Pick any length from 3 to 15 seconds, to the second, in 16:9, 9:16, 1:1, 4:3, 3:4, 21:9, or 9:21, and the clip comes back with sound generated alongside the picture.

On Treza, HappyHorse 1.1 is a model in the Video Generation node's picker, with HappyHorse 1.0 beside it. Wire reference images into the node to hold a subject steady, or wire an Image Generation node into the First frame port to animate a still. When a narrator belongs on top, a Text to Speech node and the Sequence node mix the voice over the clip's own sound. Prepaid credits, no subscription, and only successful runs are charged.

Reference-to-video

Wire in several subject or style references and HappyHorse reads the whole set, which is how a product or character carries from one shot to the next.

An exact first frame

Open on a specific image: a product still, a storyboard panel, or an Image Generation node's output wired into the First frame port.

3 to 15 seconds, to the second

Every whole-second length in that range, so a clip can be sized to the narration line it plays under instead of padded or cut to fit.

Seven formats, 21:9 to 9:21

Landscape, portrait, square, 4:3 and 3:4, plus ultrawide 21:9 and its tall 9:21 twin. One model covers a campaign's placements.

What to make with HappyHorse 1.1

Prompts, and what comes back. Each one runs on a Video Generation node on the canvas with HappyHorse 1.1 picked as the model.

“This sneaker (three product photos attached) on a wet rooftop at dusk, slow orbit”

A shot built from your product references.

“Start on this storyboard panel (attached): the door swings open onto a snowy street”

Image-to-video from an exact opening frame.

“Ultrawide 21:9 pan across a desert highway at golden hour, heat shimmer on the road”

A cinematic banner shot in one generation.

“A 9:21 rise up a glass tower at night, city lights spreading below”

An extra-tall vertical for full-screen phone placements.

“A 13 second shot of a potter centering clay, timed to the voiceover”

A take sized to the exact length of the narration.

“This mascot (references attached) waves from behind a cafe counter, square 1:1”

A feed-ready clip with your character carried over.

HappyHorse 1.1 is one node on a canvas

Chain it with language, image, and video models, then publish the whole pipeline as an API. See how a full AI video pipeline fits together.

Step 01
Draft the brief
A language model node expands a one-line idea into a detailed, on-brand prompt before it ever reaches the model.
Step 02
Generate with HappyHorse 1.1
Run variants in parallel, compare against other models on the same prompt, and keep every result in run history.
Step 03
Publish it as an API
The pipeline becomes a versioned endpoint. Your product calls it with the typed /invoke API or any OpenAI SDK.

HappyHorse 1.1 as an API, priced per clip

The HappyHorse 1.1 API

Build a pipeline with HappyHorse 1.1 on a node and publish it. It becomes a versioned HTTP endpoint: the typed /invoke route enqueues a run and returns a run id to poll, and an OpenAI-compatible route works with SDKs you already use. Agents connect over MCP the same way. Your prompt engineering ships inside the pipeline, so every caller gets your quality, and swapping the model later does not change the API your product calls.

HappyHorse 1.1 pricing

Prepaid credits, no subscription. Each clip is charged at HappyHorse 1.1's own rate from your balance, shown before you run, and only successful generations are charged, so a failed run costs nothing. Credit packs start at $5 and never expire, which suits spiky production schedules. Run history records the cost of every generation, per node, so spend is auditable rather than a surprise.

HappyHorse 1.1, answered

What is the difference between HappyHorse 1.1 and 1.0?

The spec sheet is the same: 3 to 15 second clips, seven aspect ratios, reference images, and a first frame. The difference is price: at 1080p, 1.1 costs less per second. That makes 1.1 the natural default, and comparing it against 1.0 on the same prompt is one dropdown on the node.

Can I use reference images and a first frame together?

Use one per generation. Choose references when a product or character has to stay recognizable, and a first frame when the clip has to open on an exact image. If both are set, the node uses the first frame and leaves a note that the references were skipped. For both at once, generate the opening still from your references with an Image Generation node, then animate it through the First frame port.

How does HappyHorse 1.1 compare to Wan 2.7?

Both are Alibaba models that take reference images. HappyHorse 1.1 runs longer, 3 to 15 seconds against 2 to 10, and covers seven aspect ratios to Wan 2.7's five. Wan 2.7 adds a last-frame keyframe, which makes it the pick when a clip has to land on an exact closing frame. They share one picker, so run the brief on both and keep the better take.

How much does HappyHorse 1.1 cost on Treza?

Treza uses prepaid credits with no subscription. Each clip is charged at HappyHorse 1.1's rate from your balance, and only successful generations are charged. Credit packs start at $5 and never expire.

Can I call HappyHorse 1.1 through an API?

Yes. Build a pipeline that uses HappyHorse 1.1 and publish it as a versioned HTTP endpoint. Call the typed /invoke API from your product, or point any OpenAI SDK at the OpenAI-compatible endpoint. If a better model ships later, swap it on the node without changing the API your product calls.