Features
dbt documentation is the built-in system that auto-generates searchable HTML docs for dbt projects, capturing model, source, test, and lineage metadata.
dbt documentation is the collection of HTML pages that dbt automatically builds from your project’s model code, YAML metadata, and database reflections. It centralizes lineage graphs, column descriptions, tests, and freshness info so analysts can explore your warehouse without reading SQL.
Complete docs turn tribal knowledge into discoverable assets, speeding onboarding and reducing errors. When every table and column is described, developers write SQL faster and BI users trust metrics.
Run dbt docs generate to create a target/catalog.json and target/manifest.json with model metadata, then launch a local site with dbt docs serve. Commit the target/ folder to CI or publish to S3 for team access.
# build models
$ dbt build
# generate docs
$ dbt docs generate
# serve at http://localhost:8080
$ dbt docs serve
This sequence compiles models, captures lineage, and spins up an interactive UI.
Add description fields to every model, source, and column YAML. Enforce docs coverage in CI with the dbt-expectations package. Review docs with pull-request templates that flag missing text.
Treat docs like code. Store YAML and Markdown alongside models. Use Git branches for updates, and preview docs in CI so reviewers can click through lineage before merging.
Galaxy’s SQL editor surfaces dbt model metadata inline, letting engineers hover to see column descriptions while writing queries. The AI copilot pulls docs context to autocomplete joins accurately and suggest missing filters.
Missing YAML descriptions break trust. Fix by adding CI checks that fail builds when description is null. Another issue is outdated lineage after refactors—schedule nightly dbt docs generate in Airflow or GitHub Actions.
dbt docs convert opaque SQL into transparent knowledge assets. Having a live lineage graph and searchable definitions ensures every stakeholder speaks the same language, accelerates development, and prevents costly misunderstandings.
Sync the target/ folder to S3 or deploy with Netlify. Point colleagues to the static site URL.
Yes—host the static site behind SSO or VPC rules. Alternatively, use Galaxy where access inherits database permissions.
Galaxy scans your manifest.json, mapping models to database tables so its AI copilot and autocomplete show accurate column docs.
dbt will still compile, but CI checks can fail the pipeline. Add descriptions to restore coverage and keep docs useful.
.avif){kind=logo}
