29 studio partners. 100+ licensed titles. 25 modalities across Core, Add-ons and Premium, frame-synced on one clock. Capture scoped to your spec, with Services available per order.
Explore the data, methods, and checkpoints behind our latest research. The complete release record lives on Hugging Face.
Origin Lab on Hugging FaceTrain and evaluate world models with visual, spatial, input, and telemetry streams synchronized on one frame clock.
Samples require a Hugging Face access request and are for internal evaluation only. See the repository for current terms.
Explore frozen game worlds from free-camera viewpoints for spatial reconstruction and scene understanding.
Reconstruct game-world geometry and evaluate spatial models with engine-grounded flythroughs.
Study cross-player consistency with live gameplay captured on independent machines, aligned to a shared frame clock rather than rendered from replays.
Released depth checkpoints make the Game-Depth result directly inspectable and reproducible.
Study action-conditioned gameplay with RGB, depth, normals, audio, camera, and telemetry aligned to human inputs.
Train dense depth estimators on engine-grounded imagery and study transfer from game worlds to real photographs.
Earlier releases retain their Origin Lab Data License and 90-day internal evaluation track for labs. Commercial licensing requires a direct agreement. How release terms apply
Real licensed gameplay, captured in-engine at 60 fps on one clock. Use Events, Camera, Input and Feed to inspect each signal.
From rendered pixels to engine state, inspect the signals behind a capture. One synchronized record, with Core included and optional modalities scoped to your research.
The rendered world before the HUD is layered on. Engine-level RGB preserves the scene without the player's interface.
9 core modalities in every hour, 60 fps, frame-synced on one clock, 100,000+ hours ready to ship.
Inspect the live capture aboveCore is priced per player-hour. Add-ons, premium modalities and custom capture are quoted per order. Request pricing and we'll scope it with you.
Real world models cannot be built off hyper-sensationalized 10-second clips of headshots uploaded online. You need the diversity. The boring. You need everything - not the 1% that makes it online, or even the 10% that video-game players do. You need the 90% of things they would never do, but CAN do.
This is why Origin Lab's data is different.
Signals are only as good as the hours behind them. Our data is captured by professional players following structured task lists, planned for diversity per hour, and monitored in real time. Every recording earns its place in the corpus.
Granular, per-game instructions. Complete every crafting loop. Drive backwards through every biome. Play only at night. Interact with every NPC type.
Each run is planned to cover the widest range of combinatorial actions, environments, and edge cases per hour. Models see the full breadth of what a game world can produce.
Webcam-based engagement tracking and input-activity detection. AI-driven coverage planning steers future runs toward gaps. No idle time enters the corpus.
Share your spec or RFP. We design capture plans around your requirements and execute across any title in our catalog.
Signals and sessions only matter if they reach you cleanly. The platform threads rights holders, our capture team, and your training pipeline through one license, one schema, one API.
29 studio partners. 100+ licensed titles across a wide range of game genres, plus 3D environments. Don't see your target genre, mechanic, environment, or modality? We capture to spec.
Engine-level capture from AAA and indie titles. First-party exports with action labels, camera telemetry, depth, normals, and game state.
Worldbuilder libraries with full scene graphs. USD, FBX, glTF assets ready for simulation, training, and synthetic data.
Share a spec, an RFP, or just your training goal. We design capture plans around your requirements - across any title in our catalog, or one we license for you.
A video shows what happened. Capture from the engine can also record the geometry, inputs and world state behind it. Direct signals, not estimates made from pixels.
Per-pixel depth from the renderer, rather than depth estimated after capture. The live viewer above includes an RGB and depth comparison.
Visible imagery and sound, when present. Engine depth, camera poses and player inputs are not recoverable as original engine records from pixels alone. Training rights need separate verification.
Pair the rendered world with frame-synced engine context and recorded inputs. Scope optional modalities to your research, with a license from the rights holder.
25 modalities across three tiers. Capture scoped to your spec.
Sample pack in 24 hours. Same-day API access for evals. Whether you're training a frontier model or wiring up a single eval, start with one signed agreement and one cURL.
View