Code-First Workflows
Write workflows as ordinary code in your repository, test them locally, and deploy them like any other service.
Workflows as source code
Code-First Workflows treat orchestration as software. A workflow is a function with typed inputs, steps, and outputs. It lives in your repository, goes through code review, and is deployed by the same pipeline as the rest of your services.
Because a workflow is code, developers can reuse functions, share libraries, and write unit tests. They can compose small workflows into larger ones and keep each piece easy to reason about.
What this makes possible
- Version control, branching, and review for every change to a process.
- Unit tests that run in your existing continuous integration setup.
- Shared libraries for authentication, mapping, and error handling.
- Environment-specific configuration that keeps secrets out of the repository.
Flexible when requirements change
When a process changes, a developer changes the code. There is no separate designer to keep in sync, and a diff shows exactly what changed. Teams that already have engineers building internal services find that orchestration fits naturally into their existing practice.
If your team prefers to build its own orchestration layer, RelayStack is a faster way to get the same result, because retries, scheduling, state, and tracing are already in place.
Developers who join a project can read the workflow, run it locally, and open a pull request on the first day, because the process is documented by the code itself.