dbt review

dbt is a SQL-based tool for building and managing data models. It helps data teams turn SQL transformations into modular, tested, version-controlled pipelines, with options for local and cloud development.

The product is aimed at the modeling and development work around data pipelines. The homepage lists local validation to catch issues before they reach the warehouse, plus an interactive lineage view that traces flows across projects and columns. When a model or column is renamed, dbt says it can update downstream references automatically. Stateful builds are designed to build only models that need updates, according to the homepage.

For coding, dbt lists an IDE extension for Cursor, Claude Code, Windsurf and VS Code. Its free Developer plan includes a browser-based IDE. In short, dbt is not a general-purpose analytics suite: its core job is helping teams organize, validate and maintain SQL models.

The appeal is a coherent workflow for SQL modeling: write modular transformations, validate them locally, track them in version control and inspect their dependencies. That gives a team a way to understand what may be affected when a model changes, rather than treating each SQL file as an island.

Lineage is a useful part of that picture. dbt describes a live DAG that can trace data flows across projects and columns. Automatic refactoring is another practical feature: rename a model or column and dbt says it updates downstream references. Stateful builds can avoid rebuilding models that do not need updates, according to the homepage. The IDE extension supports Cursor, Claude Code, Windsurf and VS Code.

There is also a free entry point. The pricing page lists Developer at $0 per month for one developer, with one project, a browser-based IDE, MFA and up to 3,000 successful models per month. That is enough to explore dbt’s workflow without paying for a team plan, provided the project fits those limits.

The free plan is narrow by design: one developer, one project and 3,000 successful models per month. Teams collaborating on a first project move to Starter, listed at $100 per user per month. It includes five developer seats, one project, 15,000 successful models per month and 5,000 queried metrics per month. Annual billing prices are not published. Those are meaningful limits to check against your workload before switching.

Larger organizations have to ask for a quote: Enterprise and Enterprise+ pricing is custom. Enterprise lists 30 projects; Enterprise+ lists unlimited projects. Both include 100,000 successful models and 20,000 queried metrics per month, so the published figures do not show higher allowances for those two measures on Enterprise+.

Some costs are also harder to forecast from the pricing page. dbt State has a 30-day free trial, then costs $0.094 per DATT, with custom pricing for upfront commits. dbt Wizard is pay-per-token at model-specific rates; its free usage amounts are not published. Enterprise and Enterprise+ Wizard use beyond included credits is billed as a consumption add-on.

dbt’s Developer plan is free, but it is limited to one developer seat, 3,000 successful models per month and one project. Starter costs $100 per user per month and includes five developer seats, 15,000 successful models, 5,000 queried metrics and one project. That is a jump from $0 to a paid team plan, not a simple upgrade for one developer. The pricing page does not publish annual billing prices.

Enterprise and Enterprise+ pricing is custom. Both list 100,000 successful models and 20,000 queried metrics; Enterprise includes 30 projects, while Enterprise+ includes unlimited projects. Each plan also includes monthly dbt Wizard credits: $100 for Enterprise and $200 for Enterprise+. Extra Wizard use is billed as a consumption add-on. Wizard otherwise charges per token at model-specific rates; free usage amounts are not published. dbt State has a 30-day free trial, then costs $0.094 per DATT, with custom pricing for upfront commitments. Enterprise onboarding and professional services cost extra.

The published materials describe dbt as a SQL-based modeling tool, but do not specify a migration process, supported import formats or data-export options. Before switching, inventory your existing models and projects, then confirm how your current warehouse setup and downstream dependencies map to dbt’s projects and model lineage. The plan limits matter too: Developer and Starter each list one project, so teams with multiple projects should check Enterprise pricing before planning a move.

If you leave, the published material does not explain what dbt exports or what remains available after cancellation. Get those details from dbt before relying on its IDE, lineage or automatic refactoring in a production workflow. We were not given Reddit posts or comments for this review, so we cannot substantiate claims about why users leave dbt or how difficult they found migration. Treat switching effort and user-reported pain points as unverified until you check them against current documentation and customer reports.