Orca documentation
Build, run, and govern managed AI agents.
Orca (runorca.ai) is an open runtime for managed AI agents. It runs any supported agent harness on any supported model and governs every model and tool call from outside the agent. Orca Agent Engine mirrors the Claude Managed Agents API surface, so Anthropic-compatible clients can point their base URL at an Orca Workspace.
Agent Engine
Build, connect, and run agents
Define agents as versioned registry resources, wire them to your tools and event streams, and run them on any harness in isolated sandboxes.
Start building ↗AI Gateway
Route and govern model traffic
Put a governance plane in front of your model providers and MCP servers: vaulted credentials, routing with fallback, budgets, guardrails, and an auditable record of every call.
Explore the gateway ↗Quickstart
Define an agent, start a session, and stream its events.
View ↗Deployment
Self-host the open-source engine on your own infrastructure.
View ↗CLI and SDKs
Manage every registry resource from the terminal or in code.
View ↗API reference
Every registry operation, generated from the OpenAPI spec.
View ↗