Text to Video
happyhorse_t2v
Range: 3-15s | 720p/1080p | 16:9, 9:16, 1:1 and more
Prompt only. Use this when the idea starts as a written scene with camera, action and light.
Open this routeUlazAI - AI Image & Video Tools
New video model in UlazAI
Create short AI videos from a prompt, one source image, up to 9 reference images, or an existing clip. HappyHorse-1.1 is built for 3-15 second shots in 720p or 1080p, with stronger prompt understanding, smoother motion and better identity consistency than the previous route.
Prompt used
Use the reference person as the hero of a cinematic creator-studio promo: they step into soft window light, camera pushes in, subtle hand motion, polished UGC commercial look, no text, no logos.
Model video
Use this video as the quick model check: accepted input, generated output and the clearest route in Video Studio.
Input Choose text-to-video, image-to-video, reference-to-video, or video edit from your source material.
Sample Watch generated clips before spending credits.
Route Prompt, pricing, and next step sit together in one workflow.
Choose the route from the source material. Start from text, animate one still, preserve identity with references, or edit a clip you already have.
happyhorse_t2v
Range: 3-15s | 720p/1080p | 16:9, 9:16, 1:1 and more
Prompt only. Use this when the idea starts as a written scene with camera, action and light.
Open this routehappyhorse_i2v
Range: 3-15s | one image URL | 720p/1080p
Requires one source image up to 20MB. Best when a first frame, product shot or character still should stay recognizable.
Open this routehappyhorse_r2v
Range: 3-15s | up to 9 reference images | broad aspect ratios
Use up to 9 reference images and labels such as [Image 1] when identity, style or composition consistency matters.
Open this routehappyhorse_videoedit
Range: source video URL | prompt-based edits
Requires a source video URL for targeted changes to motion, atmosphere or visual details.
Open this routeThe upgrade is strongest when a clip needs to keep a subject, product or style recognizable while the shot gains motion.
Use prompts up to 5000 non-Chinese characters. The model handles camera notes, subject action, style and reference labels such as [Image 1] more directly.
Reference-to-video supports up to 9 reference images, useful for character consistency, product shape, styling and campaign visuals.
Use 16:9, 9:16, 3:4, 4:3, 4:5, 5:4, 1:1, 9:21 or 21:9 with 720p or 1080p output.
Short enough for hooks and product shots, long enough for a mini story beat, creator promo or reference-led brand clip.
HappyHorse is billed linearly by duration and resolution, so you can estimate the cost of 3, 5, 10, or 15 second clips before generating.
720p: 33 credits per second
1080p: 44 credits per second
| Duration | 720p | 1080p |
|---|---|---|
| 3s | 99 credits | 132 credits |
| 5s | 165 credits | 220 credits |
| 10s | 330 credits | 440 credits |
| 15s | 495 credits | 660 credits |
Use reference images to keep a person, mascot or stylized character recognizable across short scenes.
Start from a product render or packshot and add camera movement, hand motion, physics detail or commercial lighting.
Generate vertical hooks, UGC-style motion, campaign visuals and social cutaways with a clear source identity.
Expose text, image and reference routes separately so users can pick the exact workflow from their assets.
Use the same Video Studio route for product interfaces and workflow automation. The model slug selects the HappyHorse route.
POST /api/v1/video-studio/generate/
{
"model_slug": "happyhorse_t2v",
"prompt": "A miniature city built from cardboard comes alive at night",
"duration": 5,
"quality": "1080p",
"aspect_ratio": "16:9"
}
GET /api/v1/video-studio/status/{job_id}/
UlazAI supports text-to-video, image-to-video, reference-to-video and video edit through separate model slugs in Video Studio.
The upgrade improves prompt understanding, motion smoothness, visual quality and consistency across text, image and reference workflows.
Reference-to-video supports up to 9 reference images. Use labels such as [Image 1] in the prompt when a specific reference should guide a specific part of the shot.
HappyHorse clips run from 3 to 15 seconds. For longer productions, create multiple shots or choose another video model with longer duration support.
Text-to-video and reference-to-video support 16:9, 9:16, 3:4, 4:3, 4:5, 5:4, 1:1, 9:21 and 21:9.
The current HappyHorse routes in UlazAI focus on video output. Use models with native audio support when generated sound or spoken dialogue is required.
Open Video Studio, choose the HappyHorse route, and create a short test video with predictable credits.