A workflow is a binding of a runbook (the deterministic Python templates that drive the process) to an entity, optionally on a schedule. A runbook may invoke external agents (Claude Managed, Copilot, …) from inside its steps when it needs LLM-driven artefact generation that doesn’t fit the deterministic capabilities surface. There are two paths to a working workflow on Ntropii:Documentation Index
Fetch the complete documentation index at: https://docs.ntropii.com/llms.txt
Use this file to discover all available pages before exploring further.
Pick a template
Start from a runbook in
ntro-runbook-templates. Today: nav-monthly, nav-monthly-journals, document-ingest, expense-processor. Configure for your tenant + entity and deploy.Build a new runbook
When no template fits, author a new runbook from scratch —
@workflow.defn class, @ui_step methods, HITL gates. The platform’s coding-agent assistant can scaffold from a skill definition.Deploying a workflow
Either path lands at the samentro workflow create --path flow. The CLI uploads the runbook bundle, binds it to an entity, and (optionally) attaches a schedule:
runbookSlug (the deployed code identifier) + entityId (where it operates) + an optional schedule. The same runbook can be deployed under multiple tenants — each deployment becomes a separate workflow.
Importing an agent
If your workflow needs a step that generates a complex artefact (auditor handover .docx, narrative summary, etc.), register an external agent and invoke it from inside a runbook step. See Register agents.Next steps
Build runbooks
Concept walkthrough:
NtroWorkflow, @ui_step, HITL gates, child workflows.Skill definitions
The runbook.md / skill.md frontmatter the platform’s coding-agent assistant uses to scaffold a workflow.
Register agents
Bring an external agent (Claude Managed, Copilot) into a runbook step.
Deploy to production
Push runbooks at versioned paths; pin versions per tenant.