Synthesis Scenarios
Craft sophisticated data scenarios for multi-human ML models.
Data Generation at Scale




Complex Multi-human Use Cases
Place multiple animated humans in a scene for advanced ML models. All humans are perfectly labeled in each frame for use cases such as localization, avatar creation, pose understanding, sub-segmentation and more.Diverse Environments
Choose the location that’s right for your use case, like outdoor or indoor, home or office. We also provide variable lighting conditions based on light source(s) and time of day.Tailored Activities
Scenarios are based on multiple humans in motion. Configure randomized and contextual activities based on where each human is located within the environment.Integrate & Iterate
Our platform is for ML practitioners first. We provide python libraries to quickly and easily retrieve all the metadata for your use case. Once you integrate, create new data on the fly and watch your models improve.Easy to Create Labeled Data



Step 1
Choose environment
Step 2
Define multi-human actions
Step 3
Render
Easy to Generate Media

Step 1
Choose environment

Step 2
Define multi-human actions

Step 3
Render
Features
Multi-human Rich Labels
Photorealistic Animation
Variable Camera Locations
Multi-human Rich Labels
Multi-human Rich Labels
Pixel-perfect set of rich labels for every human in a scene, with detailed segmentation maps, depth, surface normals, 2D/3D landmarks, and more.

Photorealistic Animation
Photorealistic Animation
A unique combination of generative neural networks, procedural generation, and cinematic VFX rendering systems deliver photorealistic images and video for complex scenarios.

Variable Camera Locations
Variable Camera Locations
Easily capture data across variable camera positions and angles. Setup custom camera intrinsics to match your hardware and use-case.

Multi-human Rich Labels
Photorealistic Animation
Variable Camera Locations
Multi-human Rich Labels
Pixel-perfect set of rich labels for every human in a scene, with detailed segmentation maps, depth, surface normals, 2D/3D landmarks, and more.

Photorealistic Animation
A unique combination of generative neural networks, procedural generation, and cinematic VFX rendering systems deliver photorealistic images and video for complex scenarios.

Variable Camera Locations
Easily capture data across variable camera positions and angles. Setup custom camera intrinsics to match your hardware and use-case.
