dbt docs combine YAML descriptions with warehouse metadata

The page states that good model documentation helps downstream consumers discover and understand curated datasets. dbt generates docs from project artifacts and from queries against the warehouse information schema, then renders them as a site. Descriptions are added in the same YAML files used for data tests.

For data and content teams, that puts discoverability next to the code that defines datasets rather than in a separate wiki. Model code, lineage DAG, tests, column types, and table sizes appear together for consumers who need context before trusting a table.

Governed shared meaning needs human-readable definitions bound to the same resources machines parse: descriptions on models and columns become the prose layer over structured project and warehouse metadata.