Skip to content

Is ETLantic for me?

Status: Available in ETLantic 0.53.0 (Beta release candidate).

ETLantic is a typed modeling, validation, and planning layer for Python data pipelines. It is not a dataframe engine, warehouse transformer, or durable orchestrator.

Quick decision

If you need… Start with Use ETLantic when…
SQL analytics projects in a warehouse dbt You also need typed multi-engine Python pipelines
Durable DAG scheduling / ops Airflow / Dagster / Prefect You want typed contracts and secret-free plans that compile or execute into those tools
Row-level dataframe checks Pandera / Great Expectations You also need graph wiring and contract validation before run
Typed pipeline composition across engines ETLantic You want one logical model, fail-closed plans, and interchangeable backends

Compared to common tools

Tool Primary job Relationship
dbt SQL transformation project / warehouse analytics Complementary
Airflow Orchestration and scheduling Complementary — etlantic-airflow compiles plans to DAG artifacts (does not install Airflow)
Prefect Orchestration and scheduling Complementary — etlantic-prefect is a local direct-execution scheduler MVP, not a DAG compiler
Dagster Orchestration and software-defined assets Complementary — compiler bridge planned for 0.55, not shipped
Pandera / GE Dataframe / table validation libraries Complementary — ETLantic validates wiring and contracts; row suites stay engine-side
Polars / Pandas / Spark / SQL engines Execution Complementary — install matching etlantic-* plugins

What ETLantic owns

  • Typed Data, Transformation, and Pipeline authoring
  • Multi-phase validation and deterministic, secret-free PipelinePlan
  • ODCS / DTCS / DPCS interchange
  • Local runtime, reports, schema observations, and plugin trust controls

What ETLantic does not own

  • Dataframe compute kernels
  • Formal enterprise SLA / unbounded shared-service multi-tenant SaaS
  • Managed Spark clusters
  • Broader compliance programs and multi-tenant attestation (adopters own these; ETLantic ships release SHA-256 digests / attestations for its own packages since 0.20; CycloneDX SBOM is optional and verified at tag time) Supported isolation profiles are Available via the control plane — see Capabilities

When NOT to use ETLantic

Prefer another tool (or use ETLantic only as a thin companion) when:

Situation Prefer
Warehouse-only SQL analytics with dbt already owning the project dbt alone
You only need a durable multi-worker scheduler / ops UI Airflow / Dagster / Prefect alone
You only need row-level dataframe checks inside notebooks Pandera / Great Expectations alone
You need unrestricted shared-service multi-tenant SaaS or a formal SLA today Wait / supplement — Supported profiles Available; community non-SLA; Operator Console is 0.56
You need bronze/silver/gold medallion vocabulary as core APIs SparkForge / medallantic, not core ETLantic
You refuse typed Python modeling and only want SQL files on disk Stick with your SQL toolchain

If several rows match, ETLantic is usually complementary — not a replacement.

ETLantic will not silently apply AI proposals, ship write MCP tools, or replace Airflow/dbt. Proposals stay untrusted until a current 0.42 approval.

Next

  1. Installation
  2. Quickstart
  3. Capabilities
  4. Evaluator brief for enterprise decision packets