The dbt workbench that never leaves your machine.

• Everything in Base
• Wicky, the local AI assistant — ask questions about your models and your data, explain failing runs and tests, and generate docs and data tests you review before saving, on a model that runs on your machine. No API key, nothing sent to the cloud. More about Wicky
• Column-level lineage tracing across the DAG, derived from compiled SQL — no yml docs required, and included in interactive DAG exports
• Column-level impact analysis — only the columns, models & tests a change reaches, including through filters and joins
• Contract checks — before you merge, the columns your changes remove, rename or retype and everything that still uses them (models, tests, exposures, semantic models, metrics, dashboards and monitors), plus models that no longer match their enforced contract
• Data diff — build your changed models in dev and compare them with prod row by row: rows added, removed and changed, which columns changed and by how much, schema differences and sample rows, in-warehouse or across separate connections
• Lineage propagation — fill empty column descriptions from the upstream columns they pass through, and trace tagged columns like pii to every model, column and exposure they reach
• Run performance insights — each run's critical path, models that got slower than usual, and recommendations like making a model incremental or raising threads
• Refactors — rename a model or a column everywhere it's used, or extract a CTE into its own model, reviewed as one diff before anything is written, with dbt parse afterwards and one-click undo
• Smarter test suggestions — relationships and not_null on the columns other models join on, unique on each model's grain, ranked by how much depends on them with the evidence for each; Health → Test Gaps adds them across many models at once, no AI model needed
• Semantic layer drafts — “Add to semantic layer” on any model: Wicky drafts the semantic model and starting metrics, checks them with MetricFlow and runs each one before you review the YAML; Query data answers with your defined metrics when they cover the question, and the metric explorer's query line takes a question in words and fills itself in
• Metric monitors — watch metrics for spikes, drops, level shifts, missing periods and restated history against a seasonal baseline; checked after builds that touch them, with status on the Health page, DAG, Impact and dashboard tiles
• Metric compare — see what your changes do to a metric's numbers before you merge: uncommitted changes vs HEAD, your branch vs its base, dev vs prod or two metrics, with the rows that moved and the SQL diff
• Metric explanations — “Explain this number” writes up what the debug checks and the definition say about a metric's value, most likely cause first
• New Pro features as they ship
• Priority support

You'll get a license key by email — paste it into Settings → Subscription in the app. How it works