Generate action scaffold code for an Orca machine in TypeScript, Python, or Go. Use when the user has a verified machine and wants implementation stubs for the action functions. When the machine file also contains decision tables, compiled evaluator functions and wired action stubs are included automatically.
Compile an Orca machine to XState v5 TypeScript config or a Mermaid state diagram. Use when the user wants to generate runnable code or a visual diagram from a .orca.md file.
Generate a coordinated multi-machine Orca system from a natural language spec. Use when the user needs multiple state machines that invoke each other (e.g. a coordinator with child workers).
Generate a verified Orca state machine from a natural language spec. Use when the user wants to design a new state machine, workflow, or agent orchestration from a description.
Verify an Orca machine for structural correctness, completeness, and determinism. Use when the user wants to check a .orca.md file for errors or warnings.
AGENTS.md instructions for jascal/orca-lang, covering orca — agent integration guide, what orca provides, installation, mcp server (recommended for modern agent frameworks) and or run directly with npx.
Go from natural language to verified finite state machines — topology bugs caught before code runs. Runs locally from the @orcalang/orca-mcp-server npm package. Needs 3 environment variables to run.