Create an interactive 3D simulation from a text prompt and explore it right in your browser
The free Odyssey-3 Flash demo generates a navigable 3D world in real time - you enter via a text description, walk around in first- or third-person, and trigger events on the go
- 1.Open the free interactive Odyssey-3 preview page at experience.odyssey.systems in your browser. The demo runs on the lightweight Odyssey-3 Flash model; no sign-up or payment required.
If the page is unavailable in your region, try refreshing the tab or coming back later - the preview is a research preview and may temporarily limit the number of concurrent sessions.
Odyssey-3 Flash Demo ↗ - 2.In the text request field, paste a prompt describing the scene you want to generate. The more specific the details of the environment, lighting, and objects, the more predictable the result.
The prompt must be in English - that is the language the model was trained on. Avoid overly long descriptions: 2 - 4 sentences are enough.
Odyssey-3 Flash Demo ↗PromptA [BIOME] landscape at [TIME_OF_DAY]. In the center: [LANDMARK_OBJECT]. Surrounding terrain: [TERRAIN_DETAILS]. Atmosphere: [WEATHER_CONDITION].
What this prompt doesAn English template for describing a 3D scene. Replace [BIOME] with a terrain type (for example, «desert», «forest», «arctic»), [TIME_OF_DAY] with the time of day («sunset», «midnight», «dawn»), [LANDMARK_OBJECT] with a key object («a crashed spaceship», «an abandoned lighthouse», «a stone bridge»), [TERRAIN_DETAILS] with terrain details («rolling hills with dry grass», «cracked salt flats», «snow-covered rocks»), [WEATHER_CONDITION] with the weather («thick fog», «clear sky», «lightning storm»). - 3.Click the generate button. The model will create an interactive 3D scene at 832x480 resolution and open it for immediate navigation. Generation happens in real time - frames are rendered as you move.
The first run may take a few seconds to load and initialize. After that, the world is generated in a stream as you move.
Odyssey-3 Flash Demo ↗ - 4.Choose a view: first-person (First-Person) for full immersion or third-person (Third-Person) to see the character from the side. The camera is controlled separately from movement - you can look around without moving.
The view switch is usually available in the demo control panel. If the panel is not visible right away, hover over the bottom or top edge of the screen.
Odyssey-3 Flash Demo ↗ - 5.Move around the generated world using the keyboard (WASD) and mouse. During generation, you can trigger events - for example, add an object or change a condition - and watch the model compute the scene's reaction in real time.
Physics is calculated on the fly, not pre-rendered. If you encounter artifacts or «crumbling» textures, this is a limitation of the research preview - the model is still being improved.
Odyssey-3 Flash Demo ↗ - 6.If you are a developer and want to integrate world generation into your project, apply for Odyssey-3 API access through developer.odyssey.ml. In the application, describe your use case and team.
API pricing has not been announced yet, and latency metrics and SLAs have not been published. Treating the model as production-ready is only possible as an experiment for now.
Odyssey-3 API ↗
- freeOdyssey-3 Flash Demo (browser preview)Free demo with no sign-up or payment; session limits and the duration of free access are not specified by the company.
- APIOdyssey-3 API (for developers)Access by request via developer.odyssey.ml; pricing, limits, and SLAs have not been published as of the announcement.
Why today
AI services change fast - interfaces and free limits may differ from what's described.