Pipelines and automation¶
This route is for automation and data-tool builders who want typed boundaries, reproducible commands, and CI-friendly native artifacts.
Choose your route¶
First projectBuild a staged workflowCompose an executable project with explicit inputs, steps, failures, outputs, and tests.
OperateDiagnose a failed buildCapture structured diagnostics, explain a stable code, repair the source, and inspect the backend.
Preview · DeliveryPrepare a CI artifact laneUse the verified local gate while hosted 0.5 toolchain installation remains a packaging preview.
OperateUse the CI contractsApply formatting, testing, locked/offline, exit-code, and report surfaces in automation.
The typed data processor is the executable next step when your workflow reads typed JSON, validates records, and emits a report artifact.
Install once¶
Follow Getting Started for the current installer choices and canonical first-project commands.
What transfers¶
- Staged transformations, explicit inputs and outputs, and repeatable operational boundaries
- Deterministic commands, exit codes, locked dependencies, and inspectable reports
- Modules, file I/O, typed errors, tests, and release artifacts
What changes¶
- Each fallible boundary returns a typed value rather than relying on logs or ambient shell state.
- Data contracts and transformations can live in ordinary program modules with tests.
- A native executable becomes the repeatable automation unit.
What not to expect¶
- SQL, dataframe, Spark, or notebook compatibility; classify relational query and plan work with the Incan or IncQL? chooser
- A remote package registry or hosted pipeline service
- Enterprise deployment policy beyond the documented lock, offline, report, and exit-code contracts