Skip to content

PhysicalAI API

Preview Notice
The PhysicalAI API is currently in preview and not yet available for use. This feature is under active development and will be released soon. Examples, code, and additional documentation will also be made available upon release.

The PhysicalAI API provides advanced tools for video-based understanding, simulation, and synthesis of the physical world. It supports intelligent video processing using pixel-level analytics, semantic search, and generative modeling.

Capabilities

  • Scene Understanding: Analyze physical scenes with spatial and temporal awareness.
  • Object Localization: Detect and track objects throughout video frames.
  • Key Event Retrieval: Identify and timestamp important actions and anomalies.
  • Pixel-Level Segmentation: Overlay or isolate specific regions using semantic segmentation.
  • Content Synthesis: Generate synthetic video clips based on prompts or video references.
  • Semantic Video Search: Query indexed videos using natural language.