A curated Python wrapper and prompt library for the HappyHorse 1.0 API (developed by Alibaba's Taotian Group), delivered via muapi.ai. Generate cinematic AI videos from text prompts and static images — currently the #1 ranked AI video generation model — at native 1080p HD or the cheaper 720p tier (~half price), and use the bundled prompt pack of high-performing community examples to get great output on the first try.

Join subreddit HappyHorseAI for discussion.

Try for free ArenaAI.

🌊 Also explore these top AI video models:

• 🎬 Seedance 2.0 API — ByteDance's cinematic 2K model with character sheets, omni-reference & video edit
• 🎥 Veo 4 API — Google DeepMind's native 4K model with audio, character consistency & camera controls

• 💳 Pro or Business plan required. Free-plan users get 403 { "detail": "Happy Horse 1.0 is available only on the Pro and Business plans. Upgrade at https://muapi.ai/topup#plans." }
• 🎨 Use Happy Horse today directly inside the muapi playground on a Pro ($20/mo) or Business ($100/mo) plan. Early access to new models is a paid-plan perk.

The Python SDK and MCP server in this repo mirror the final public surface so you can wire them up now; they'll start succeeding the moment the API goes GA.

• Why HappyHorse 1.0
• Key Features
• Installation
• MCP Server
• Quick Start (Python)
• API Endpoints & Reference
• API Method Reference
• Pricing
• Prompt Library ← popular community prompts
• Official Resources
• License

HappyHorse 1.0 is Alibaba's state-of-the-art AI video generation model, built by the Future Life Lab team at Taotian Group. It debuted anonymously on April 7, 2026, instantly claiming the top spot in both Text-to-Video and Image-to-Video public benchmarks.

• #1 Ranked: 1333 Elo in T2V, 1392 Elo in I2V — surpassing every competitor in public benchmarks.
• Native 1080p HD: Full HD output without upscaling, powered by a 15B-parameter 40-layer Transformer architecture.
• 720p tier: same model, ~half the cost — pick resolution="720p" when you don't need full HD.
• Blazing Fast: ~10 seconds average generation time, one of the fastest available models.
• Developer-First: Simple Python SDK via MuAPI infrastructure.

• ✅ Text-to-Video (T2V) — 1080p — POST /api/v1/happy-horse-1-text-to-video-1080p.
• ✅ Text-to-Video (T2V) — 720p — POST /api/v1/happy-horse-1-text-to-video-720p (~half the 1080p cost).
• ✅ Image-to-Video (I2V) — 1080p — POST /api/v1/happy-horse-1-image-to-video-1080p.
• ✅ Image-to-Video (I2V) — 720p — POST /api/v1/happy-horse-1-image-to-video-720p (~half the 1080p cost).
• ✅ Reference-to-Video (Ref2V) — 1080p / 720p — POST /api/v1/happy-horse-1-reference-to-video-{1080p|720p}. Generate a video from a text prompt + 1–9 reference images (style, subject, environment) — different from I2V's single start frame.
• ✅ Video Edit — 1080p / 720p — POST /api/v1/happy-horse-1-video-edit-{1080p|720p}. Edit an existing video with a natural-language instruction; optionally anchor 0–5 reference images, keep or regenerate the audio track.
• ✅ Flexible Aspect Ratios: 16:9, 9:16 (TikTok/Reels), 1:1, 4:3, 3:4.
• ✅ Duration: 4–15 seconds per clip.
• ✅ Two output tiers: native 1080p HD or budget-friendly 720p.

pip install happyhorse-1-api

# Clone the HappyHorse 1.0 API repository
git clone https://github.com/Anil-matcha/Awesome-HappyHorse-1.0-API-and-Prompt.git
cd Awesome-HappyHorse-1.0-API-and-Prompt

# Install required dependencies
pip install -r requirements.txt

Create a .env file in the root directory and add your MuAPI API key:

MUAPI_API_KEY=your_muapi_api_key_here

Use HappyHorse 1.0 as an MCP (Model Context Protocol) server, allowing AI assistants like Claude Desktop or Cursor to directly invoke HappyHorse generation tools.

1. Ensure MUAPI_API_KEY is set in your environment.
2. Run the server:

   python3 mcp_server.py

3. To test with the MCP Inspector:

   npx -y @modelcontextprotocol/inspector python3 mcp_server.py

from happyhorse_api import HappyHorseAPI

# Initialize the HappyHorse 1.0 client
api = HappyHorseAPI()

# 1. Generate Video from Text (T2V) — 1080p (default)
print("Generating AI Video using HappyHorse 1.0...")
submission = api.text_to_video(
    prompt="A cinematic aerial shot of a coastal city at golden hour, waves crashing against cliffs, birds flying",
    aspect_ratio="16:9",
    duration=10,
    # resolution="720p",   # uncomment to halve the cost at 720p
)

# 2. Wait for completion
result = api.wait_for_completion(submission['request_id'])
print(f"Success! View your HappyHorse 1.0 video here: {result['outputs'][0]}")

# 3. Reference-to-Video — generate from a prompt + up to 9 reference images
ref_submission = api.reference_to_video(
    prompt="The same character walks through a rainy neon-lit Tokyo alley at night",
    images_list=[
        "https://example.com/character_front.jpg",
        "https://example.com/character_profile.jpg",
        "https://example.com/style_neon_alley.jpg",
    ],
    aspect_ratio="9:16",
    duration=8,
)
ref_result = api.wait_for_completion(ref_submission["request_id"])
print(f"Reference-to-video output: {ref_result['outputs'][0]}")

# 4. Video Edit — edit an existing video with a natural-language instruction
edit_submission = api.video_edit(
    prompt="Replace the daytime sky with a stormy sunset and add lightning flashes",
    video_url="https://example.com/source_clip.mp4",
    audio_setting="origin",   # keep the original audio; use "auto" to regenerate
)
edit_result = api.wait_for_completion(edit_submission["request_id"])
print(f"Video-edit output: {edit_result['outputs'][0]}")

Base URL: https://api.muapi.ai/api/v1 Auth: x-api-key: <YOUR_API_KEY> (API-key access is currently in closed beta — see the note at the top of the README)

All responses return {"request_id": "<id>", "status": "processing"}; poll GET /api/v1/predictions/{request_id}/result until status is completed and read outputs[0] for the final muapi-hosted video URL.

Endpoint: POST https://api.muapi.ai/api/v1/happy-horse-1-text-to-video-1080p

curl --location --request POST "https://api.muapi.ai/api/v1/happy-horse-1-text-to-video-1080p" \
  --header "Content-Type: application/json" \
  --header "x-api-key: YOUR_API_KEY" \
  --data-raw '{
      "prompt": "A majestic eagle soaring over snow-capped mountains at sunrise",
      "aspect_ratio": "16:9",
      "duration": 10
  }'

Endpoint: POST https://api.muapi.ai/api/v1/happy-horse-1-text-to-video-720p

Identical request body to the 1080p endpoint above. Output resolution is encoded in the URL — costs ~half of the 1080p tier (see Pricing).

curl --location --request POST "https://api.muapi.ai/api/v1/happy-horse-1-text-to-video-720p" \
  --header "Content-Type: application/json" \
  --header "x-api-key: YOUR_API_KEY" \
  --data-raw '{
      "prompt": "A majestic eagle soaring over snow-capped mountains at sunrise",
      "aspect_ratio": "16:9",
      "duration": 10
  }'

Endpoint: POST https://api.muapi.ai/api/v1/happy-horse-1-image-to-video-1080p

The first image in images_list is used as the start frame and the generated clip animates outward from it.

curl --location --request POST "https://api.muapi.ai/api/v1/happy-horse-1-image-to-video-1080p" \
  --header "Content-Type: application/json" \
  --header "x-api-key: YOUR_API_KEY" \
  --data-raw '{
      "prompt": "The clouds drift slowly, light shifts from golden to dusk",
      "images_list": ["https://example.com/landscape.jpg"],
      "aspect_ratio": "16:9",
      "duration": 10
  }'

Endpoint: POST https://api.muapi.ai/api/v1/happy-horse-1-image-to-video-720p

Same body as the 1080p I2V endpoint; cheaper 720p output.

curl --location --request POST "https://api.muapi.ai/api/v1/happy-horse-1-image-to-video-720p" \
  --header "Content-Type: application/json" \
  --header "x-api-key: YOUR_API_KEY" \
  --data-raw '{
      "prompt": "The clouds drift slowly, light shifts from golden to dusk",
      "images_list": ["https://example.com/landscape.jpg"],
      "aspect_ratio": "16:9",
      "duration": 10
  }'

Endpoint: POST https://api.muapi.ai/api/v1/happy-horse-1-reference-to-video-1080p

Generate a video from a text prompt and 1–9 reference images. Each image is treated as a style/subject reference (character, environment, look) — this is different from I2V, which uses a single image as the start frame.

curl --location --request POST "https://api.muapi.ai/api/v1/happy-horse-1-reference-to-video-1080p" \
  --header "Content-Type: application/json" \
  --header "x-api-key: YOUR_API_KEY" \
  --data-raw '{
      "prompt": "The same character walks through a rainy neon-lit Tokyo alley at night",
      "images_list": [
          "https://example.com/character_front.jpg",
          "https://example.com/character_profile.jpg",
          "https://example.com/style_neon_alley.jpg"
      ],
      "aspect_ratio": "9:16",
      "duration": 8
  }'

Endpoint: POST https://api.muapi.ai/api/v1/happy-horse-1-reference-to-video-720p

Identical request body to the 1080p reference-to-video endpoint; cheaper 720p output.

curl --location --request POST "https://api.muapi.ai/api/v1/happy-horse-1-reference-to-video-720p" \
  --header "Content-Type: application/json" \
  --header "x-api-key: YOUR_API_KEY" \
  --data-raw '{
      "prompt": "The same character walks through a rainy neon-lit Tokyo alley at night",
      "images_list": [
          "https://example.com/character_front.jpg",
          "https://example.com/style_neon_alley.jpg"
      ],
      "aspect_ratio": "9:16",
      "duration": 8
  }'

Endpoint: POST https://api.muapi.ai/api/v1/happy-horse-1-video-edit-1080p

Edit an existing video with a natural-language instruction. Optionally supply 0–5 reference images to anchor the look/characters/objects that should appear in the edited output, and choose whether to regenerate the audio (auto) or keep the original audio track (origin).

curl --location --request POST "https://api.muapi.ai/api/v1/happy-horse-1-video-edit-1080p" \
  --header "Content-Type: application/json" \
  --header "x-api-key: YOUR_API_KEY" \
  --data-raw '{
      "prompt": "Replace the daytime sky with a stormy sunset and add lightning flashes",
      "video_url": "https://example.com/source_clip.mp4",
      "audio_setting": "origin"
  }'

Endpoint: POST https://api.muapi.ai/api/v1/happy-horse-1-video-edit-720p

Same body as the 1080p video-edit endpoint; cheaper 720p output.

curl --location --request POST "https://api.muapi.ai/api/v1/happy-horse-1-video-edit-720p" \
  --header "Content-Type: application/json" \
  --header "x-api-key: YOUR_API_KEY" \
  --data-raw '{
      "prompt": "Restyle the entire clip as a 1990s anime cel-shaded look",
      "video_url": "https://example.com/source_clip.mp4",
      "images_list": ["https://example.com/anime_style_ref.jpg"],
      "audio_setting": "auto"
  }'

Endpoint: GET https://api.muapi.ai/api/v1/predictions/{request_id}/result

Returns the muapi standard prediction envelope — status transitions through queued → processing → completed (or failed). On completion the outputs array contains muapi-hosted video URLs; the top-level video key mirrors outputs[0] for convenience.

Endpoint: POST https://api.muapi.ai/api/v1/upload_file

Multipart-form helper to host a local image on muapi so it can be referenced in images_list.

HappyHorse 1.0 bills flat per-second. 720p costs half of 1080p across every endpoint. Reference-to-Video and Video Edit price slightly above plain T2V/I2V because they consume more reference / video conditioning.

Pricing shown here matches the live cost-strategy on muapi.

A curated pack of high-performing prompts for HappyHorse 1.0, organized by use case. Drop any of these directly into text_to_video(...) or the /happy-horse-1-text-to-video-{1080p,720p} endpoints.

Cave Flashlight Cinematic

A flashlight beam exploring a cave system, illuminating wet limestone formations. The light catches crystalline calcite deposits that glitter and flash. Where the beam passes through shallow standing water, it creates bright caustic patterns on the submerged floor. Stalactites cast long, swinging shadows as the flashlight moves.

Flower Time-Lapse Continuity

A flower blooming and wilting over two weeks, one photo per day. Same vase, same window, same angle. Light changes with weather.

Tracking Shot Street Escape

TRACKING SHOT follows her from behind as she runs through the street. Sari fabric flows and trails behind her, catching the wind. CLOSE-UP on bare feet hitting the ground. Fabric billowing. She glances back. Keeps running. Determined.

Voice Assistant Day-In-The-Life

A time-lapse of a family using a home voice assistant throughout the day. From setting morning alarms, playing music, checking the weather, and controlling smart lights, the product integrates seamlessly into their daily life.

1990s Action Cartoon Firebending

1990s action cartoon style. A young martial artist performs a firebending kata. The flames are hand-drawn with thick outlines and bold orange-yellow gradients. Dynamic camera swoops around the character. The fighting stance shows anime influence while maintaining Western animation proportions. Smoke effects use the signature layered look of the era.

Cyberpunk Android Repair Bay

Cyberpunk anime style. A female android sits in a maintenance chair as robotic arms repair her damaged arm. The skin panel is open, revealing intricate servos and fiber-optic cables beneath. Her eyes are blank and unfocused during the repair cycle. Neon city lights filter through rain-streaked windows. Cool blue and pink color palette with high contrast shadows.

Graduation Banner Chaos

A massive "CONGRATULATIONS GRADUATES" banner being unfurled across a university building by maintenance workers on the roof. The wind catches it mid-unfurl, turning it into a sail that nearly lifts one worker off their feet. Coworkers grab him, everyone laughs, and the banner finally drops into place. Below, students start taking selfies immediately.

• Name the shot. Tokens like TRACKING SHOT, CLOSE-UP, WIDE ANGLE, TIME-LAPSE meaningfully change camera behavior.
• Specify the style bucket. 1990s action cartoon style, cyberpunk anime style, cinematic 35mm film — style tokens front-load the aesthetic.
• Keep motion concrete. "Wind catches it mid-unfurl" beats "it moves dramatically."
• For I2V, describe the motion, not the subject. The image already provides the subject — the prompt should describe what changes.

• API Provider: MuAPI.ai — get your MUAPI_API_KEY and access the playground here
• Upgrade plan: muapi.ai/topup#plans — Pro ($20/mo) or Business ($100/mo) unlocks early access to new models including Happy Horse 1.0

MIT — see the LICENSE file.

Keywords: HappyHorse 1.0 API, Awesome HappyHorse 1.0, HappyHorse 1.0 Prompts, Alibaba HappyHorse, AI Video Generator, Text-to-Video AI, Image-to-Video API, Reference-to-Video API, Video Edit API, HappyHorse Python SDK, Alibaba Video AI, MuAPI, Video Generation API, Native 1080p AI Video, AI Video Creation, HappyHorse API Documentation, HappyHorse I2V, HappyHorse T2V, HappyHorse Ref2V, HappyHorse Video Edit, AI Movie Generator, Python Video API, HappyHorse Tutorial, #1 AI Video Model.