Puppi

Puppi runs data pipelines you write in TypeScript. You declare assets with source, model, and destination. The CLI bundles that project and hands it to the engine. The engine runs the graph and records each result as a materialization.

An asset has a key, such as core/customer, and a function that produces its rows. A model names what it reads with ctx.input("demo/customers"). That call is the dependency. Lineage, staleness, and quarantine are all read from the recorded results.

The sidebar switches between three sections. This one is the documentation. The knowledge base is the task list for a run that went wrong. The API is the HTTP control plane.

Start here

• Quickstart writes a small pipeline and runs it.
• Concepts cover the bundle, deploys, materializations, lineage, staleness, and quarantine.
• Connectors are the clients a pipeline uses to call an API.
• CLI and configuration are the reference.