Skip to content

VideoOdyssey

Odyssey

Real-time world model that generates interactive, explorable AI video.

Categories
VideoGaming
Pricing
FREE
Source
Proprietary
Hosting
Cloud
Platforms
Web
Models
Self-contained (on-device)
Verified
Jun 15, 2026

Odyssey is a world-model startup whose AI generates interactive video you can steer in real time — explorable, 3D-like scenes streamed frame by frame without a game engine. Its current model, Odyssey-2, turns a text prompt or image into a multi-minute simulation you navigate as it is generated. It is positioned for film, gaming, and interactive media, and is available as a free research-preview experience.

Capabilities 3

What it actually does — grouped by capability family.

  • Text-to-video (secondary capability)
  • Image-to-video (secondary capability)
  • World model (primary capability)

Pros & cons

  • Real-time interactive generation
  • Steerable, multi-minute scenes
  • No game engine or assets needed
  • Free research-preview access
  • Research preview, not production-ready
  • Limited resolution / visual fidelity
  • No commercial terms or pricing yet

Tags

Further reading

Alternatives to Odyssey (4)

View all Video →
  • View Decart details
    GamingPAID

    Decart

    Decart

    Real-time generative world and video models (Oasis, Lucy).

    Decart is an AI lab building real-time generative world and video models, including Oasis (a fully AI-generated, playable Minecraft-style open world) and Lucy (real-time video transformation used in gaming, ads, and virtual try-on). Its Oasis 3 is positioned as an API-accessible world model for Physical AI, with a low-latency inference stack. Products are accessed via the platform API and interactive demos.

    Real-time world-model generation
    Sales-led, API-first, no free tier
    • gaming
    • world-models
    • real-time-video
    • generative
    Open
  • View Marble details
    3DFREEMIUM

    Marble

    World Labs

    Generate editable 3D worlds from text, images, or video.

    A generative world model from World Labs that turns text prompts, photos, videos, panoramas, or 3D layouts into editable, downloadable 3D environments. Outputs can be refined and exported for use in game engines and 3D pipelines.

    Exports to game engines and 3D pipelines
    Quality degrades away from start point
    • world-model
    • 3d-generation
    • spatial
    • scene-generation
    Open
  • View Runway details
    VideoFREEMIUM

    Runway

    Runway

    Video AI with Gen-series models and a full timeline editor.

    End-to-end video AI platform — text-to-video, image-to-video, in-painting, motion brush, and a timeline editor. Longest-running studio in the space; default choice when the output ships to clients.

    Full editor, not just a generator
    Credits deplete fast on long clips
    • video-gen
    • editing
    • motion
    • film
    Open
  • View Luma Dream Machine details
    VideoFREEMIUM

    Luma Dream Machine

    Luma AI

    Mobile-first AI video generation from text and images.

    Luma Labs' video model surfaced as an iOS app and web tool. Strong text-to-video, image-to-video, and keyframe controls — the indie creator's go-to before reaching for Runway's heavier toolkit.

    Fast generations
    Less granular control than Runway
    • video-gen
    • image-to-video
    • keyframes
    • mobile
    Open
Open Odyssey