Happy OysterThe Open-Ended AI World Model

The world is your oyster. Open it. Happy Oyster is an open-ended world model for real-time world creation and interaction. Direct a running scene or wander an infinitely extendable world using text, voice, or images, with joint audio-video generation and coherent physics over time. Built by the Alibaba team behind HappyHorse-1.0, available now on fal as an official API partner.

What makes Happy Oyster different

Real-Time World Model

Generation and Interaction at Once

Happy Oyster does not render a finished clip and hand it back. It generates the world as you steer it, responding to your text, voice, and image input live. You direct the scene while it is being created, so the model reacts in the moment instead of after the fact.

Joint Audio-Video

Sound, Vision, and Physics Together

The model produces joint audio and video in a single continuous stream, with physics that stay coherent over time. Objects keep their place, light behaves consistently, and motion carries through as the world evolves around you.

Open-Ended Worlds

Worlds That Keep Growing

Start from a single prompt and keep going. In Wandering mode the environment extends dynamically as you explore in first person, so the world grows around you instead of ending at the edge of a frame.

Two modes

Direct a scene or wander a world

Happy Oyster runs in two modes. Directing puts you in control of a live scene as it plays. Wandering drops you inside a world you can walk through. Both take multimodal input and return joint audio and video.

Steer a running world

Directing

Direct a live scene as it plays. Change the setting, the characters, and the camera angle in real time while lighting, gravity, and motion stay consistent across the shot.

Length
Up to 3 minutes
Resolution
480p or 720p
Control
Text, voice, or images
Explore in first person

Wandering

Drop into a world from a single prompt and move through it freely in first person. The environment keeps extending as you go, growing around you instead of ending at the edge of a frame.

Length
Up to 1 minute
Resolution
480p
Control
WASD and camera controls
API Documentation

How to access the Happy Oyster API

The client API handles the request submit protocol. It handles the request status updates and returns the result when the request is completed.

import { fal } from "@fal-ai/client";

const result = await fal.subscribe("alibaba/happy-oyster/text-to-world", {
  input: {
    prompt: "A neon-lit night market on a rainy street, camera drifting between the stalls.",
    resolution: "720p",
    duration: 10,
  },
  logs: true,
  onQueueUpdate: (update) => {
    if (update.status === "IN_PROGRESS") {
      update.logs.map((log) => log.message).forEach(console.log);
    }
  },
});

console.log(result.data);
console.log(result.requestId);
Examples

See what Happy Oyster can create

Worlds directed and wandered with Happy Oyster, generated in real time with joint audio and video.

Getting Started

Happy Oyster API Integration Steps

Get up and running in minutes. No GPUs to manage, no infrastructure to set up.

1.  Install the client

Pick your package manager. For Python, use pip.
npm install --save @fal-ai/client

2.  Create an account on fal

Sign up to get access to the dashboard and your API keys.

3.  Get your API key

Locate your API credentials in the developer dashboard. Set FAL_KEY as an environment variable in your runtime.

4.  Submit a request

Use fal.subscribe() to submit your request with a prompt and parameters. The client handles the async queue automatically and returns the final result when generation is complete.
Try it now

No setup required

Start building Happy Oyster worlds instantly in the playground. No API key needed, just describe your world and hit generate.

Open Playground
For developers

Integrate via API

Grab an API key from your dashboard and integrate Happy Oyster into your app with a few lines of code. Python and JavaScript SDKs available, plus a REST API for any language.

API Documentation
For teams

Scale in production

Run Happy Oyster at scale with dedicated capacity and enterprise support. Talk to our team about volume pricing and deployment.

Contact Sales
FAQ

Common questions about Happy Oyster

What is Happy Oyster?

Happy Oyster is Alibaba's open-ended AI world model. Instead of returning a finished video clip, it builds an interactive 3D world in real time and responds to you as it generates, taking text, voice, image, and keyboard input and producing joint audio-video output with coherent physics.

What is the difference between Directing and Wandering?

Directing lets you steer a running scene in real time for up to three minutes at 480p or 720p using text, voice, or images. Wandering lets you move through an endlessly expanding world in first person for up to one minute at 480p using standard WASD and camera controls.

How is Happy Oyster different from a video generator?

A video generator produces a finished clip after you prompt it. Happy Oyster generates and responds at the same time, so you shape the world while it is being created. Lighting, gravity, and motion stay consistent as the scene evolves rather than being fixed once at the end.

Who built Happy Oyster?

Happy Oyster comes from the Alibaba team behind HappyHorse-1.0, the Taotian Future Life Lab under Alibaba Token Hub, led by Zhang Di. It was unveiled in April 2026 as Alibaba's entry into real-time AI world models.

Can I use Happy Oyster right now?

Yes. Happy Oyster is available now on fal via both the playground and the API. Text-to-world, image-to-world, and exploration endpoints are live at /models/alibaba/happy-oyster/text-to-world, /models/alibaba/happy-oyster/image-to-world, and /models/alibaba/happy-oyster/explore.

What can I build with Happy Oyster?

Happy Oyster is built for real-time 3D content: game prototyping and level ideation, film and virtual-production previsualization, interactive experiences, and training or simulation environments. Anywhere you want to explore a world rather than watch a fixed clip.

Does Happy Oyster generate audio?

Yes. Happy Oyster performs joint audio-video generation, producing sound and picture together in a single continuous stream rather than adding audio in a separate post-processing step.

How much does Happy Oyster cost?

Happy Oyster uses pay-per-second pricing on fal, billed by the second of generated world with no minimums or subscriptions. See the model page for current per-second rates, and contact sales for enterprise pricing.

How do I access Happy Oyster?

Happy Oyster is available through the official API on fal. Sign up, grab an API key from your dashboard, and call the text-to-world, image-to-world, or exploration endpoints, or try it first in the playground with no setup.

Ready to transform your enterprise with AI?

Take the first step towards AI-driven innovation. Our team of ML engineers is ready to help you prototype, develop, and scale your AI solutions.

Enterprise Contact Form