Scripted Systems

While Nebula has full SDKs for TypeScript and C# to build systems you deploy anywhere in the cloud, Scripted Systems let you write behaviors directly from your application and deploy them instantly. No infrastructure to manage, no deploy pipelines to configure.

What Are Scripted Systems?

Nebula provides a scripted worker pool — a managed runtime that holds all your scripts and manages their lifecycle. Combined with the system editor (an in-app code editor powered by @nebula/react), you can write, test, and deploy system logic without leaving your application.

The system editor integrates directly into your Nebula app through the React SDK. It provides:

A Monaco-based code editor with TypeScript support
Instant deployment — save and your changes are live
Full revision history for every system
Live log viewing from running system instances

When to Use Scripted vs. External

Not every system needs a full deployment pipeline. The choice between scripted and external depends on your workload:

	Scripted Systems	External SDK
Best for	Business rules, data transforms, lightweight automation	Heavy computation, complex dependencies, custom environments
Deploy	Instant from the in-app editor	Your own CI/CD pipeline
Runtime	Managed by Nebula's worker pool	Your own infrastructure
Language	TypeScript	TypeScript or C#
Iteration speed	Seconds	Minutes to hours

Scripted Systems will become the recommended way to develop Nebula application behaviors. The external SDKs will remain fully supported for heavier workloads where you need more control over the environment. Nebula will eventually auto-scale the scripted worker pool based on system queue demand.

The Development Loop

Building with scripted systems is a tight feedback loop:

1. Open the system editor — The editor is a React component from @nebula/react that you embed in your app's admin interface. It connects directly to your Nebula instance.

2. Write your behavior — Define a query and processing logic using the same patterns as external systems. The editor provides autocomplete, type checking, and access to your component library.

3. Deploy instantly — Click deploy and your system is live. The scripted worker pool picks up the new version immediately. No build step, no container image, no waiting.

4. View live logs — The editor shows real-time logs from your running system. See exactly which entities are being processed, what changes are being made, and any errors that occur.

5. Iterate — Make changes and redeploy. Every revision is saved, so you can compare versions and roll back if needed. The feedback loop is measured in seconds, not deployment cycles.

This workflow is particularly valuable for business rules that change frequently. A denial prioritization formula, a team assignment algorithm, a notification trigger — these are the kinds of logic that benefit from rapid iteration without the overhead of a full deployment.

Scripted Systems use the same Behavior model and participate in the same Orchestration pipeline as external systems. The difference is only in how you deploy them — the runtime guarantees are identical.