video model · by Alibaba
HappyHorse 1.0 AI Video Generator
Generate reference-driven AI videos with up to 9 reference images using HappyHorse 1.0 - Alibaba's multi-reference video model. Strong visual guidance, text-to-video, image-to-video, and video editing in a single model, now inside Fliki.
Generated with HappyHorse 1.0
A handful of HappyHorse 1.0 clips generated inside Fliki. No edits, no post.
Prompt
Prompt
Prompt
Prompt
Prompt
Prompt
Prompt
Trusted by 50,000+ companies worldwide
Why HappyHorse 1.0 stands apart
Up to 9 reference images
HappyHorse 1.0 accepts up to 9 visual reference images. Use them to lock characters, objects, styling, and scene elements at the same time — far beyond single-reference models.
Reference-to-video workflow
HappyHorse is built around reference guidance — give it the visuals you want preserved, and the prompt does the rest. It's the right model when consistency across assets matters.
Text-to-video and image-to-video
Run from a pure text prompt for conceptual work, or anchor with a first-frame image for locked visuals. Both modes inherit reference-image guidance.
Video editing input
HappyHorse accepts a single video input for editing workflows — modify, restyle, or extend an existing clip rather than generate from scratch.
Up to 15-second clips
Generate clips from 3 to 15 seconds. Long enough for full narrative beats, short enough to keep iteration fast on volume work.
Multi-aspect support
Compose natively at 16:9, 9:16, or 1:1. The model lays the frame for the chosen ratio rather than cropping a master.
Strong identity retention
Multi-reference design lets HappyHorse hold character, product, and styling identities across the full clip — useful when an asset needs to look unmistakably 'on brand'.
Run inside Fliki
No separate Alibaba account, no Runware setup. HappyHorse runs through the standard Fliki prompt interface — upload references, write a prompt, generate.
Native lip-sync in seven languages
HappyHorse 1.0 supports native lip-sync across Mandarin, Cantonese, English, Japanese, Korean, German, and French — useful for multilingual presenter content and localized brand work.
How it works
How to generate a video with HappyHorse 1.0
HappyHorse runs inside Fliki's playground. Six steps.
Write your prompt
HappyHorse responds well to descriptive prompts. Describe the subject, action, and the style cue you want the model to lean into.
Select HappyHorse 1.0 as your model
Pick HappyHorse 1.0 from Fliki's playground model selector. Your prompt routes to the Alibaba model via Runware.
Add reference images
Upload up to 9 reference images. HappyHorse uses them as visual guidance — character, style, object, and scene anchors. The more references, the stronger the guidance.
Pick aspect ratio and duration
Choose 16:9, 9:16, or 1:1. Set duration anywhere from 3 to 15 seconds — useful for short hooks all the way to full beats.
Optionally add a video for editing
HappyHorse supports a single video input for editing workflows. Use it when you want the model to modify an existing clip rather than generate from scratch.
Generate at 720p
Hit Generate. HappyHorse outputs a 720p clip with multi-reference guidance baked in.
HappyHorse FAQ
Frequently asked questions
Everything you need to know about generating with HappyHorse 1.0 inside Fliki.
What is HappyHorse 1.0?
HappyHorse 1.0 is Alibaba's reference-to-video model, accessible on Fliki via Runware. It supports text-to-video, image-to-video, and multi-reference workflows — accepting up to 9 visual reference images for guidance.
How many reference images can HappyHorse use?
Up to 9 visual reference images, plus a separate first-frame image for image-to-video. The multi-reference design is the model’s biggest differentiator.
What inputs does HappyHorse accept?
Text prompts, a first-frame image (image-to-video), up to 9 visual reference images for style and identity guidance, and a single video input for editing workflows.
What resolution does HappyHorse output?
HappyHorse 1.0 generates at 720p across 16:9, 9:16, and 1:1 aspect ratios.
How long can HappyHorse videos be?
Clips can run from 3 to 15 seconds. The longer durations carry full narrative beats; shorter durations are best for volume iteration.
Does HappyHorse generate audio?
HappyHorse focuses on video and reference fidelity. For audio-first workflows, Veo 3.1 Fast, Seedance 2.0, or Kling 3.0 Pro are stronger picks — all available on Fliki.
Still curious?
Try Fliki free in your browser, no credit card required.
Start freeCan HappyHorse edit existing video?
Yes. HappyHorse supports a single video input for editing workflows — restyle, modify, or extend an existing clip with prompt and reference guidance.
When should I pick HappyHorse over other models?
When you need consistent identity across multiple visual elements — characters, products, styling — and a single reference image isn’t enough. The 9-reference ceiling is what makes HappyHorse distinct.
AI video models
Discover more
Tools
Discover features
Generate reference-driven video with HappyHorse.
Up to 9 reference images for character, style, and scene guidance. Free to start, no credit card required.
Generate your first video freeFree forever plan · No credit card required · Cancel anytime