dbt is a transformation framework for the "T" in ELT. Analysts define models as version-controlled SQL (with Jinja templating); dbt compiles them into a dependency graph and executes them inside the target warehouse or engine — it brings software-engineering practice to SQL rather than moving data itself.
The framework includes schema and data tests, generated documentation, and model-level lineage. Incremental models, snapshots, and packages cover common warehouse transformation patterns.
dbt Core is open source and CLI-driven; dbt Cloud adds a managed scheduler, IDE, and APIs. It has become the default transformation layer over warehouses such as Snowflake and BigQuery, usually orchestrated by tools like Airflow or Dagster.