How It Works
World Simulator is an immersive, AI-powered text-adventure game that brings your choices to life through video, in first person perspective.
NOTE: This is not real-time. It takes ≈30 sec. to generate each video step.
Powered by OpenAI's Sora 2, World Simulator generates video sequences that respond to your actions. Each decision you make influences the next scene, creating a unique, personalized adventure.
- Cinematic Immersion: Every step of your journey is visualized through AI-generated video
- Interactive Storytelling: Text-adventure gameplay enhanced with visual feedback
- Dynamic World: Your choices shape the story and influence subsequent video generations
- Flexible Duration: Choose between 4, 8, or 12-second video clips per action
How It Works
Step-by-Step Experience
1. Choose Your World
Select from a variety of pre-made scenarios or create your own. Each scenario sets the stage for your adventure with unique settings, characters, and narratives.
2. Watch the Opening Scene
Your adventure begins with an AI-generated video that establishes your character's position and the environment. This immersive introduction sets the tone for your journey.
3. Enter Your Action
Like in classic text-adventure games, type what you want to do. Be descriptive! For example:
- "I walk through the ancient stone archway into the forest"
- "I approach the mysterious stranger at the bar"
- "I examine the glowing artifact on the pedestal"
4. AI Generates Your Next Scene
Based on your action and the story so far, the game generates a new video showing the result of your choice, mantaining narrative continuity while responding to your decisions.
5. Continue Your Adventure
Each action leads to a new video, building your unique story. The experience continues as long as you want to explore, with the AI adapting to your choices.
Video Generation
Video Duration Options
Choose the length of video that best suits your storytelling pace. Each duration offers a different balance between detail and credit cost.
| Duration | Credit Cost | Best For |
|---|---|---|
| 4 seconds | 10 credits | Quick actions, rapid exploration |
| 8 seconds | 20 credits | Standard scenes, balanced pacing |
| 12 seconds | 30 credits | Cinematic moments, detailed sequences |
Generation Time
Video generation takes approximately 30 seconds per step. While not real-time yet, this allows the AI to create high-quality, contextually relevant videos that maintain story coherence.
Memory & Continuity
Memory system helps maintain visual and narrative consistency between scenes. The AI remembers your character, the environment, and key story elements to create a cohesive experience.
Credits & Pricing
Credit-Based System
World Simulator uses a credit system to power video generation. Credits are consumed each time you generate a new video step.
Pricing Options
Platform Credits
Purchase credits directly through World Simulator to use across all your adventures.
- 4-second video: 10 credits per step
- 8-second video: 20 credits per step
- 12-second video: 30 credits per step
Use Your Own OpenAI API Key
Advanced users can connect their own OpenAI API key to use World Simulator. This option requires:
- An OpenAI account with organization verification
- Access to the Sora 2 API
- Sufficient balance in your OpenAI account
When using your own API key, you'll be billed directly through your OpenAI account according to Sora 2's current pricing. View the latest rates at openai.com/api/pricing.
🔄 Automatic Refunds
Credits are automatically refunded when generation fails due to safety filters, technical errors, or other issues. You only pay for successful video generations.
Current Limitations
World Simulator is an experimental platform pushing the boundaries of AI-powered interactive storytelling. Please be aware of these current limitations:
⚠️ Safety Filters
Many steps may fail due to OpenAI's safety filters. When this happens, the credits used for failed steps are automatically refunded to your account. This is a temporary limitation as we work to improve content generation reliability.
⚠️ Visual Coherence
While the memory system maintains continuity, some scenes may show artifacts or inconsistencies that don't align perfectly with the story. Characters, objects, or environments may occasionally appear different between scenes. We're continuously working to improve coherence.
Generation Time
Current video generation takes approximately 30 seconds per step. We're working toward a real-time experience as the technology evolves.
Future Improvements
We're actively developing solutions to address these limitations, including:
- Enhanced safety filter handling
- Improved scene-to-scene coherence
- Faster generation times
- Better character and object consistency