Skip to main content
pragma-os lets you define AI agents, models, tools, and cloud infrastructure as YAML resources. Apply them with a single command, and the platform handles provisioning, dependency resolution, and automatic change propagation.

​
Get started

Deploy Your First AI Agent

Zero to a running AI agent in 20 minutes. Create a secret, configure a model, define an agent, and deploy it.

Build a Reactive AI Pipeline

Multi-agent teams with tools, knowledge bases, and reactive dependencies. 45 minutes.

Create a Custom Provider

Extend pragma-os with your own resource types. 60 minutes.

​
Why pragma-os

Reactive dependencies — Resources reference each other. When an upstream resource changes, dependents rebuild automatically. Swap a model, and every agent using it redeploys. Provider model — Providers encapsulate lifecycle logic for any service. The platform ships with providers for GCP, Kubernetes, and Agno. Build your own in Python. Real-time visibility — A web UI shows your dependency graph, resource states, and live event feeds. Watch changes propagate through your infrastructure.

​
Learn more

How It Works

Understand the pragma-os model.

CLI Reference

Complete command reference.

Resource Catalog

Browse available providers and resource types.

Build a Provider

Create custom providers for any service.