HappyHorse 1.1 AI Video Generator
Make cinematic multi-shot videos with dialogue and sound from a prompt or a photo, using Alibaba's HappyHorse 1.1 and HappyHorse 1.0.
Why Make HappyHorse Videos on getimg.ai
HappyHorse 1.1 and 1.0 are ready the moment you sign in, with no API keys, separate accounts or minimum spend.
Scenes written like a script
Describe the light, the action, the lines people say and how the camera moves. HappyHorse turns that into a multi-shot clip with film-like texture and mood.
Start from your own images
Use a photo as the first frame, or add reference images of a person, product or place. HappyHorse keeps them looking the same from shot to shot.
Commercial rights on every paid plan
Paid plans include commercial usage rights for your HappyHorse videos, subject to our terms. Publish them, use them for clients or run them as ads.
How to Use HappyHorse on getimg.ai
Your first HappyHorse clip takes three steps.
1. Open Create video
Click here and select HappyHorse 1.1 or HappyHorse 1.0 from the model picker in the prompt box.
2. Write a prompt
Describe the scene, the action and the sound. See our video prompting guide and the example on the right. If you like, add a first frame or reference images.
3. Generate
Pick 16:9, 9:16, 1:1, 4:3 or 3:4, a length up to 15 seconds and a resolution. Video and sound come back together in one clip.
A dimly lit Victorian study filled with leather-bound books and flickering candlelight. A middle-aged man in a velvet waistcoat is frantically searching through a drawer. A woman enters the frame from the shadows, her face half-hidden by a lace veil. The camera cuts from a medium shot to a tight close-up of her lips as she whispers: “It’s not in the drawer, Arthur. I moved it weeks ago.” He freezes, looking up with wide, panicked eyes: “Where is it?” Dramatic chiaroscuro lighting, shallow depth of field, rich fabric textures, dust motes dancing in the light, 19th-century oil painting aesthetic, suspenseful atmosphere.
How to Prompt HappyHorse
HappyHorse reads a prompt like a short script. Write it shot by shot: "Shot 1", "Shot 2", "Shot 3", with a framing for each one, such as a wide shot, a close-up or a medium shot. The model cuts where you put the breaks.
Give each shot one action. A beekeeper lifting a frame, then honey dripping, then a smile and a line of dialogue reads cleanly. Three actions crammed into one shot don't.
Put spoken lines in quotes and say who speaks them, then end with a "Sound:" line for ambience and effects. Camera and light go last: handheld, slow push-in, warm low sun, shallow depth of field.
Short prompts work too. Prompt enhancement fills in detail when you start from a single line.
Shot 1: Wide shot of a beekeeper in a white veil lifting a wooden frame out of a hive in a lavender field at sunrise. Shot 2: Close-up of the frame, golden honey dripping from the comb, bees crawling over it. Shot 3: Medium shot as she lifts the veil, smiles, and says: "Best harvest in years." Handheld camera, warm low sun, shallow depth of field. Sound: buzzing bees, birdsong, soft wind through the lavender.
HappyHorse Prompt Templates
Copy a template, swap the parts in brackets and generate. Each one plays to something HappyHorse does well.
Multi-shot scene
Shot 1: Wide shot of [a person] [doing something] in [a place], [time of day]. Shot 2: Close-up of [a detail of the action]. Shot 3: Medium shot as [the person] looks up and says: "[a short line]." Handheld camera, [light], shallow depth of field. Sound: [two or three background sounds].
Keep it to three or four shots for a 10 to 15 second clip, so each shot has time to play.
1080p Video with a Film Look
Shallow depth of field, soft falloff, grain and texture in skin, fabric and concrete: HappyHorse is tuned for the look of footage shot on a real camera.
HappyHorse 1.1 renders at 480p, 720p or 1080p, and HappyHorse 1.0 at 720p or 1080p. Draft at a lower resolution, then render the version you keep at 1080p.
A Snorricam shot of a man sprinting across a gravel-covered rooftop. The background of the city skyline shakes violently while his face remains stable. Shot 2: A cut to a wide profile shot as he leaps across a massive gap between two buildings. The wind whips his jacket realistically. Shot 3: He lands hard on the other side, rolling into a crouch. He looks back and says: “That’s the only way out.” High-speed cinematography, handheld look, 16mm film grain, realistic physics, intense facial performance, cold urban color palette.
Multi-Shot Scenes That Stay Consistent
One prompt can hold several shots. HappyHorse handles the cuts, the camera moves and the pacing, so a clip plays like an edited scene instead of a single take.
People and objects keep their look and position across the cuts, which matters for short dramas, product reveals and dialogue scenes.
A 16:9 top-down shot of a cardboard shoe box being opened. Tissue paper crinkles realistically as it's moved away to show the shoe. Shot 2: A person lifts a clean white leather sneaker out of the box, turning it to catch the light on the metallic logo. Shot 3: The person holds the shoe up near their face and says: "The stitching on these is actually perfect." Handheld, slightly shaky camera, realistic fabric and leather textures, natural indoor lighting, 35mm film grain, modern streetwear aesthetic.
Audio Generated with the Video
Sound comes out in the same pass as the picture: dialogue lip-synced to the speaker, ambience that fits the setting, effects and music.
Write the sound into the prompt like the rest of the scene. Put spoken lines in quotes, then list the background sounds. Leave the sound out and HappyHorse picks it for you.
Shot 1: A close-up of a pianist’s hands playing a complex, melancholic jazz chord on a vintage upright piano. Audio: The deep, wooden resonance of the piano notes and the soft mechanical "thud" of the keys being pressed. Shot 2: A cut to a medium shot of a woman in a velvet dress leaning against the piano, a microphone in front of her. She begins to sing a soulful, low-register jazz melody: "I wish you would stay...". Audio: Her voice is smoky and rich, perfectly lip-synced to her mouth movements. Shot 3: A wide shot showing the club. Audio: The piano and vocals continue, layered over the ambient "clink" of glasses and hushed crowd chatter. Style: 35mm film, moody shadows, warm amber lighting, sophisticated noir aesthetic.
HappyHorse 1.1 vs HappyHorse 1.0
We ran the same prompt through both versions at 1080p, 8 seconds, 16:9, with sound on. The clip plays the HappyHorse 1.1 result first, then the HappyHorse 1.0 result, each labeled.
HappyHorse 1.1
Up to 15 seconds at 480p, 720p or 1080p, with up to 9 reference images or a first frame, and native audio. In our test it rendered sharper detail, kept her face steady through the line and followed the tracking shot closely.
HappyHorse 1.0
Up to 15 seconds at 720p or 1080p, with up to 4 reference images or a first frame, and native audio. In our test it gave a warmer, grainier film grade and a tighter push-in, and it uses fewer credits per clip.
Which one to pick
Use HappyHorse 1.1 for final renders, scenes with several reference images and quick 480p drafts. Keep HappyHorse 1.0 for its warmer film look or for projects you started on it.
A young woman with a bleached buzz cut, freckles and gold hoop earrings skateboards down an empty concrete parking ramp at golden hour. She carves a wide turn, kicks the board up into her hand, looks at the camera and says: "Okay, one more." Tracking shot alongside her, then a slow push-in on her face. Warm low sun, long shadows on the concrete. Sound: wheels rumbling on concrete, wind, distant city traffic.
Explore more AI video models and tools
Frequently Asked Questions
HappyHorse is a family of AI video models from Alibaba. It turns text prompts and images into cinematic video clips with native audio, and it's known for multi-shot scenes where people and objects stay consistent across cuts.
On getimg.ai you can use HappyHorse 1.1, the current version, and HappyHorse 1.0, the original release, from the same model picker in Create video.
HappyHorse 1.1 is the newer version. It adds 480p for quick drafts, takes up to 9 reference images instead of 4, and in our tests renders sharper detail and steadier faces during dialogue. HappyHorse 1.0 makes clips at 720p or 1080p, has a warmer, grainier film look and uses fewer credits per clip. Both make clips up to 15 seconds with native audio, in 16:9, 9:16, 1:1, 4:3 and 3:4.
Open the Create video generator, click the model button in the prompt box and pick HappyHorse 1.1 or HappyHorse 1.0. Choose an aspect ratio (16:9, 9:16, 1:1, 4:3 or 3:4), a length and a resolution, write your prompt and generate.
For better results, describe the subject, the action, the setting and the sound, and split longer scenes into shots. Our video prompting guide has more examples.
Yes. Upload a photo as the first frame and describe the motion, the camera move and the sound, and HappyHorse animates it through getimg.ai's image to video feature. You can also add reference images instead, so a person or product appears in a new scene without fixing the opening shot.
Both versions make clips up to 15 seconds. Longer clips leave room for three or four shots, while 5 to 8 seconds suits a single shot or a quick reveal.
Yes. HappyHorse generates dialogue, ambience, effects and music together with the video, in a single pass. Write spoken lines in quotes in your prompt so the model knows what each person says.
Yes. You can access HappyHorse 1.1 through the getimg.ai API for programmatic video generation, and HappyHorse 1.0 is in the API too. They sit alongside other leading video models in the API, including Kling and Seedance.
The API runs on its own credit balance, separate from your getimg.ai app subscription. You top it up as needed and pay only for the videos you generate.
Paid plans include commercial usage rights for eligible generated content, subject to our terms and applicable third-party rights. Uploaded references and third-party intellectual property may still carry their own restrictions.
Make Your First HappyHorse Video
Start from a prompt or a photo and get a multi-shot clip with dialogue and sound, up to 15 seconds at 1080p.