Structural Causal Model Visualization

OpenCM 1.0

The Universal Interchange Standard for Causal AI.

Causal AI — and We’re Giving It Away

Read the SpecExplore Models

The Infrastructure Gap in Causal AI

The year is 2026. Causal AI is finally having its moment. After decades of Judea Pearl’s pioneering work on causal inference and structural causal models (SCMs), enterprises are waking up to a fundamental truth: correlation is cheap, causation is priceless.

But causal inference has no portable model format. OpenCM does for causal inference what ONNX did for neural networks: it makes causal models portable, versionable, composable, and transparent.

Model Registries

Maintain validated causal models—Porter’s Five Forces, PESTLE, and more. Browse the library →

The Lensing Engine

Treat models like interchangeable reasoning overlays. Swap causal worldviews as easy as changing camera lenses. See the spec →

Standardized JSON

Language-agnostic specification for variables, edges, structural equations, and explicit assumptions. See examples →