Use Case: dbt

SQL Data Lineage for dbt Models

Column-level lineage across your staging, intermediate, and mart models — using compiled SQL, no project compilation required.

dbt's docs site shows model-level lineage but not column-level lineage. When stakeholders ask 'where does mart.revenue actually come from?', model-level is not enough.

Compile once, paste forever

Run dbt compile, grab the .sql files from target/compiled, and paste them in. SQLise resolves cross-model references like normal SQL.

Layer-aware coloring

Different schemas (raw, staging, intermediate, mart) get distinct colors automatically, so the layering is visible at a glance.

Find the orphan models

Nodes with no downstream edges are likely unused — useful for cleanup PRs.

What it looks like

raw_app.usersstg_usersint_user_activitymart_active_users
raw_app sources → staging → intermediate → marts.

Most dbt projects fit comfortably in Architect (2,000 objects per parse). Enterprise covers the largest mono-repos.

Frequently asked

Do I need access to the dbt project to use this?

No. You only need the compiled SQL. You can paste a single model or all of them.

Does it understand ref() and source()?

It works on the compiled output, where ref() and source() have already been resolved into qualified table names.

Visualize your dbt models

Paste compiled SQL → see the layers.

Related