SPINE → Interactive Demo

Pipeline Builder

Build and visualize sequential processing chains. Watch data flow through each step.

Status: Reference pattern. The production compiled execution path uses SkillCompiler → PlanArtifact → PlanExecutor to run sequential steps on RunContext; this demo visualizes the generic pipeline() shape and predates the compiled path. Typed Step I/O and parallelism remain deferred.

Pipeline Steps

Pipeline Flow

Try a Preset Pipeline

How Pipelines Work

Pipelines chain steps sequentially, passing each step's output as the next step's input. If any step fails, the pipeline stops (unless continue_on_failure is set).

pipeline = Pipeline([
  Step("fetch", fetch_data),
  Step("parse", parse_json),
  Step("transform", apply_rules),
])
result = pipeline.run(initial_input)

Each step receives the previous step's output. This enables modular, composable processing chains.