Expressive Text-to-Video
Describe a scene, action, or narrative and generate a stylized clip with fluid character motion and coherent visual storytelling.
HappyHorse 1.0 is Alibaba's open-source AI video model for expressive, stylized generation and editing. It combines native 1080p output, vivid character motion, reference-conditioned scenes, and natural-language video edits across eight Muapi variants.
Closed playground beta
HappyHorse 1.0 is currently available inside the Muapi Playground on the Pro and Business plans. API-key access will open in a later public release.
HappyHorse 1.0 is Alibaba's open-source AI video model for expressive, stylized video generation. It translates text prompts, still images, and visual references into scenes with vivid character motion and dynamic storytelling.
Describe a scene, action, or narrative and generate a stylized clip with fluid character motion and coherent visual storytelling.
Bring a still image to life with optional motion direction, camera movement, and mood guidance.
Use one to nine reference images to anchor a character, product, or visual style across generated shots.
Modify wardrobe, props, environments, or character details in an existing video while preserving source motion.
Use 720p for fast, cost-efficient iteration and 1080p when the final delivery needs higher resolution.
Keep the same Muapi authentication and asynchronous request lifecycle while switching between generation and editing modes.
HappyHorse 1.0 is positioned as a leading open-source video model for text-to-video and image-to-video generation. Its combination of expressive motion, broad visual styles, native 1080p output, and fast iteration makes it useful for storyboards, social content, product animation, and game or film concept work.
All eight HappyHorse 1.0 variants use Muapi's shared API-key authentication, asynchronous task lifecycle, and prediction result endpoint. Start in the Playground to test prompts, then move the same payload shape into your application when API access is enabled for your plan. For background and release context, read the HappyHorse 1.1 API guide.
| Variant | Mode | Output | Input | Price |
|---|---|---|---|---|
| happy-horse-1-text-to-video-1080p | Text to Video | 1080p | prompt | $0.36/sec |
| happy-horse-1-text-to-video-720p | Text to Video | 720p | prompt | $0.18/sec |
| happy-horse-1-image-to-video-1080p | Image to Video | 1080p | images_list + optional prompt | $0.36/sec |
| happy-horse-1-image-to-video-720p | Image to Video | 720p | images_list + prompt | $0.18/sec |
| happy-horse-1-reference-to-video-1080p | Reference to Video | 1080p | 1–9 images + prompt | $0.42/sec |
| happy-horse-1-reference-to-video-720p | Reference to Video | 720p | 1–9 images + prompt | $0.21/sec |
| happy-horse-1-video-edit-1080p | Video Edit | 1080p | video_url + prompt | $0.42/sec |
| happy-horse-1-video-edit-720p | Video Edit | 720p | video_url + prompt | $0.21/sec |
The Playground beta currently gates access by plan. Per-second prices above reflect the current Muapi model pricing; verify the live schema and plan availability before production use.
Create short narrative shots, storyboard sequences, and character moments from written scene descriptions.
Animate product stills or reference images into on-brand clips for campaigns, ecommerce, and social media.
Turn concept art, moodboards, and visual references into moving pitch material and world-building tests.
Use natural-language edits to test wardrobe, props, weather, lighting, and environment changes without rebuilding footage.
Choose a variant, submit its schema-defined JSON payload with your Muapi API key, save the request ID, and poll the standard result endpoint. Use public image or video URLs for media inputs.
curl -X POST https://api.muapi.ai/api/v1/happy-horse-1-text-to-video-1080p \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"A playful fox runs through a sunlit meadow","aspect_ratio":"16:9","duration":5}'
# Response: {"request_id":"REQUEST_ID"}Send a prompt, aspect ratio, and duration to the 720p or 1080p text-to-video endpoint.
curl -X POST https://api.muapi.ai/api/v1/happy-horse-1-image-to-video-1080p \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"A gentle breeze moves the grass","images_list":["https://example.com/reference.jpg"],"aspect_ratio":"16:9","duration":5}'
# Response: {"request_id":"REQUEST_ID"}Pass one starting image in images_list and optionally describe the desired motion with a prompt.
curl -X POST https://api.muapi.ai/api/v1/happy-horse-1-reference-to-video-1080p \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"The character walks toward camera","images_list":["https://example.com/character.jpg"],"aspect_ratio":"16:9","duration":5}'
# Response: {"request_id":"REQUEST_ID"}Anchor a scene with one to nine reference images and a prompt describing motion, framing, and mood.
curl -X POST https://api.muapi.ai/api/v1/happy-horse-1-video-edit-1080p \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"Change the jacket to bright red","video_url":"https://example.com/source.mp4","audio_setting":"origin"}'
# Response: {"request_id":"REQUEST_ID"}Provide a video_url and natural-language edit instruction; optional reference images and audio_setting refine the result.
HappyHorse 1.0 is Alibaba's open-source AI video model family for expressive text-to-video, image-to-video, reference-to-video, and natural-language video editing.
Yes, it is available in the Muapi Playground beta on Pro and Business plans. Public API-key access is planned for a later release.
Each generation and editing mode has a 720p and a native 1080p variant. Use 720p for iteration and 1080p for higher-resolution delivery.
The reference-to-video variants accept one to nine reference images. Video-edit variants accept up to five optional reference images.
Pass one image in images_list. The prompt is optional for the 1080p image-to-video variant and can guide camera movement, motion direction, or mood.
Yes. The video-edit variants accept a video_url and a natural-language instruction, with optional reference images and audio_setting controls.
Current Muapi pricing is $0.18/sec for 720p text/image-to-video and $0.36/sec for 1080p; reference-to-video and video-edit are $0.21/sec at 720p and $0.42/sec at 1080p. Check the live Playground and plan gate before production use.
Use the closed Playground beta to test expressive video generation, reference conditioning, and natural-language editing.