HappyHorse 1.0 API
Alibaba's cinematic AI video model.
Generate, animate, edit, and extend video with Alibaba HappyHorse 1.0 — text-to-video, image-to-video, reference-to-video, video edit, and video extend, all behind one Hypereal API key. Built by Alibaba's AI lab for cinematic motion and strong prompt fidelity.
Hypereal is an independent third-party API aggregator. We are not affiliated with, endorsed by, or sponsored by Google, OpenAI, Anthropic, xAI, Black Forest Labs, ByteDance, Kuaishou, or any other model provider. Model names are trademarks of their respective owners and are used here solely to indicate which third-party model each endpoint forwards requests to.
Skatiet, ko varat izveidot
Cinematic text-to-video
“Neo-noir Tokyo street at night, neon reflections on wet pavement, slow dolly-in handheld camera”
Image-to-video animation
“Animate a portrait — subtle facial expression, natural breathing, soft hair motion”
Reference-to-video (character consistency)
“Maintain character identity and outfit across the entire fashion-walk shot”
Prompt-driven video edit
“Restyle a clip into rainy neo-noir without breaking the underlying motion”
HappyHorse 1.0 pricing
All five modes are priced the same — switch freely between them
Integrējiet dažu minūšu laikā
Standarta REST API, kas darbojas ar jebkuru valodu. Viena API atslēga sniedz piekļuvi visiem modeļiem.
- Viens galapunkts visiem modeļiem
- Bearer token autentifikācija
- JSON pieprasījuma un atbildes formāts
- Webhook atzvani asinhroniem uzdevumiem
- Pieejams Python un Node.js SDK
# HappyHorse 1.0 — Alibaba cinematic text-to-video
curl -X POST https://api.hypereal.cloud/v1/videos/generate \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "happyhorse-1-0-t2v",
"input": {
"prompt": "neon-lit Tokyo alley, slow dolly-in, cinematic handheld",
"aspect_ratio": "16:9",
"resolution": "720p",
"duration": 5
}
}'Kāpēc HappyHorse 1.0
Five modes in one model family
Text-to-video, image-to-video, reference-to-video, video edit, and video extend — all under Alibaba HappyHorse 1.0. Switch between them by changing one model slug.
720p and 1080p output
720p starts at $1.40 per 5 seconds for fast iteration. 1080p is billed at 2x the 720p rate for final delivery. Predictable cost step-up.
Strong prompt fidelity
Alibaba's training emphasizes cinematic motion, subject consistency, and prompt adherence — especially good for character continuity across reference-to-video and edit/extend chains.
Unified Hypereal API
Same /v1/videos/generate endpoint as Seedance, Kling, Veo, Sora 2, and WAN 2.7. Switch model slugs to A/B compare without re-integrating.
Kuri kredīti tiek patērēti?
Viena API atslēga darbojas abiem. Maršrutēšanu nosaka modelis, ko izsaucat, nevis atslēga.
Claude Opus 4.7, Sonnet 4.6, GPT-5.5, Gemini 3.5 Thinking un Gemini 3.5 Fast vispirms patērē Koda kredītus, pēc tam pāriet uz Vispārējiem kredītiem, ja Koda kredīti beidzas.
Attēlu, video, audio, 3D un visi citi LLM patērē tikai Vispārējos kredītus. Koda kredīti paliek rezervēti kodēšanas slodzēm.
Biežāk uzdotie jautājumi
What is HappyHorse 1.0?
HappyHorse 1.0 is Alibaba's cinematic AI video generation model, available on Hypereal in five modes: text-to-video, image-to-video, reference-to-video, video edit, and video extend.
Which HappyHorse 1.0 modes are available?
All five: T2V (text-to-video), I2V (image-to-video), Ref2V (reference-to-video), video edit (prompt-driven restyle), and video extend (continue an existing clip). They all share the same /v1/videos/generate endpoint.
How is HappyHorse 1.0 priced?
All modes cost the same: $1.40 per 5-second 720p video. 1080p is billed at 2x the 720p rate ($2.80 per 5s). Pricing scales linearly with duration (10s costs 2x a 5s clip).
Can I use multiple reference images?
Yes. happyhorse-1-0-ref2v accepts an images array of reference URLs for character or style consistency. The video-edit endpoint also accepts optional reference images for stronger visual guidance.
Can I edit or extend an existing video?
Yes. Use happyhorse-1-0-video-edit to restyle or transform a source video, and happyhorse-1-0-video-extend to seamlessly continue from the last frame of an existing clip.
How does HappyHorse compare to Seedance, Kling, or Veo?
HappyHorse 1.0 is strong for cinematic motion, character/outfit consistency, and the unique edit + extend modes in the same model family. Seedance 2.0 is faster and cheaper for high-volume T2V/I2V. Kling 3.0 has Pro/Standard tiers. Veo 3.1 has native audio. All four are accessible on Hypereal through the same /v1/videos/generate endpoint.
Is there a free trial?
Yes. Sign up and buy credits to test HappyHorse and every other Hypereal video model. Credits start at $19.99.
Build with HappyHorse 1.0 today
One API key for HappyHorse and every other Hypereal video model.