Getting started

Three on-ramps, in increasing order of infrastructure. Pick the one that matches what you have in front of you right now.

1. Installation

Add the etl facade crate to your project, choose connector features, and check the toolchain requirements (MSRV 1.94, edition 2024). Two minutes.

2. Quickstart

Run a complete pipeline — source, deserializer, operator chain, sharded sink, checkpointed commits — entirely in memory. No Kafka, no ClickHouse, no Docker. This is the fastest way to see the assembly pattern every real pipeline follows, and it doubles as the pattern for testing your own pipelines.

3. Your first pipeline

The flagship shape end to end: Kafka → Confluent-framed Avro → an operator chain → sharded ClickHouse, configured by YAML, drained gracefully on SIGTERM. Requires Kafka, a schema registry, and ClickHouse (or Docker to run them).

After this section

• The mental model behind what you just ran: Concepts, starting with Architecture.
• Every builder step in depth: Assembling a pipeline.
• Connector-specific configuration: Connectors.