dbt Core

Open-source SQL transformation framework for analytics engineers.

Reviewed by 7wData

On this page

Publisher review

dbt Core is an open-source SQL transformation framework released under Apache 2.0, acquired in 2016 and now part of the merged Fivetran-dbt Labs company (as of October 2025). It enables analytics engineers to build, test, and document data transformations using SQL and Git workflows in cloud data warehouses including Snowflake, BigQuery, Redshift, and Databricks. dbt Core handles only the "T" in ELT—transformation—requiring separate tools for data extraction, loading, and orchestration. Teams pair dbt with Airflow, Dagster, or Prefect to schedule runs, and with Fivetran or Airbyte to move data.

The framework is modular: SQL models become versioned, tested assets with automatic documentation and lineage. Custom tests catch data quality issues before analytics consumption. Macros and a package manager (over 5,000 community packages on dbt Hub) extend functionality. dbt is SQL-first, making it accessible to analysts without programming background, but projects require Git discipline, YAML configuration, Jinja templating, and CLI proficiency.

Larger repos (thousands of models) can slow CI without state management. Post-merger, dbt Labs publicly committed to maintaining Core under its current license, but the data community notes that investment may skew toward dbt Fusion (their proprietary runtime engine) and dbt Cloud (SaaS). For teams with SQL-fluent staff, self-management appetite, and separate orchestration already in place, dbt Core remains a de facto standard contract between analytics and data engineering. For smaller teams lacking Git expertise or those needing end-to-end pipeline management, the overhead outweighs the benefits.

Get the AI & data signal, daily.

335k+ subscribers read this every morning. One email, both newsletters. Unsubscribe anytime.

How it works

  1. SQL Models

    Define transformations in version-controlled SQL SELECT statements organized as a DAG; each model becomes a materialized table or view in your warehouse.

  2. Testing Framework

    Built-in tests (unique, not null, accepted values, relationships) run automatically after each build; custom SQL tests let you enforce business rules.

  3. Auto-Documentation

    Generate living documentation tied to code; changes propagate automatically, preventing drift between docs and implementation.

  4. Jinja Templating & Macros

    Use control flow, variables, and reusable SQL macros to reduce duplication and enable dynamic behavior across models.

  5. Package Manager

    Install third-party packages (utils, pre-built models, dbt configurations) from dbt Hub to extend functionality without writing custom code.

  6. Lineage & Metadata

    Automatic DAG visualization shows upstream and downstream dependencies; schema and code lineage available through dbt Cloud or artifact inspection.

  7. State-Aware Execution

    Use dbt state to run only modified models or downstream impacts, cutting CI time and warehouse costs on large projects.

Strengths and trade-offs

Strengths

  • Free and open-source under Apache 2.0; no licensing surprises or consumption-based cost multipliers.
  • SQL accessibility: teams comfortable with SQL build production transformation pipelines without Python; Git workflows enforce code review discipline.
  • Modular, testable, documented transformations become organizational assets; testing framework prevents silent data quality failures upstream.

Trade-offs

  • Handles only transformation (the 'T' in ELT); requires separate tools for ingestion, loading, and scheduling, fragmenting stack complexity.
  • 37% of users report missing functionality; no built-in streaming, no real-time, batch-only, steep learning curve (Jinja, YAML, CLI) for non-technical analysts.
  • Post-merger (Oct 2025): dbt Labs now owned by Fivetran; while committed to keeping Core open, community fears investment will shift toward dbt Fusion and Cloud, leaving Core with maintenance-only patches.

Pricing context

dbt Core is free and open-source under Apache 2.0, with no commercial licensing fees. The core SQL transformation framework, CLI, testing, documentation, and package manager are all zero-cost. Costs come from your cloud data warehouse compute (Snowflake, BigQuery, etc.) and third-party orchestrators or hosted schedulers (Airflow, Dagster, Prefect). dbt Cloud, the managed SaaS offering, uses consumption-based pricing; historical price increases of 160–1700% have driven some teams to evaluate alternatives. dbt Core remains unlimited for self-hosted deployments.

Getting started with dbt Core

  1. Install dbt Core

    Install dbt Core via pip (pip install dbt-core), then install your warehouse-specific adapter: dbt-snowflake, dbt-bigquery, dbt-redshift, or dbt-databricks. Verify installation by running dbt --version. Check that the CLI is accessible from your project directory.

  2. Connect to your warehouse

    Create a profiles.yml file with connection credentials for your cloud data warehouse. Include host, username, password, database, and schema details required by your adapter. Run dbt debug to verify connectivity. Resolve any credential or network errors before proceeding.

  3. Create dbt project and models

    Run dbt init to generate your project structure with dbt_project.yml. Write SQL SELECT statements in the /models directory for each transformation. Define tests via YAML files: unique, not null, relationships. Commit all files to Git for version control.

  4. Run dbt build

    From your project root, run dbt build to compile SQL models and materialize them as tables or views in your warehouse. Tests execute automatically after each model. Review logs for compilation errors or test failures and fix issues before moving to production.

  5. Schedule with an orchestrator

    Choose Airflow, Dagster, or Prefect to schedule dbt build runs on a production cadence. Configure your orchestrator to invoke dbt commands and monitor job completion, test results, and warehouse costs. For teams without orchestration infrastructure, dbt Cloud offers managed scheduling.

Frequently Asked Questions

What is dbt Core and what does it do?

dbt Core is an open-source SQL transformation framework released under Apache 2.0. It lets analytics engineers build, test, and document data transformations using SQL and Git workflows in cloud warehouses like Snowflake, BigQuery, and Redshift. It handles only the 'T' in ELT.

Is dbt Core free to use?

Yes, dbt Core is free and open-source under Apache 2.0 with no commercial licensing fees. The core SQL transformation framework, CLI, testing, documentation, and package manager are all zero-cost. Costs come from your cloud data warehouse compute and third-party orchestrators.

What are the main features of dbt Core?

Key features include SQL models organized as a DAG, built-in testing for data quality, auto-documentation tied to code, Jinja templating and macros for dynamic SQL, a package manager with over 5,000 community packages, lineage visualization, and state-aware execution to run only modified models.

What are the strengths and weaknesses of dbt Core?

Strengths: free and open-source, SQL accessibility for analysts, modular and testable transformations. Weaknesses: handles only transformation, no built-in scheduling or streaming, steep learning curve for non-technical users, and 37% of users report missing functionality.

How does the Fivetran-dbt Labs merger affect dbt Core?

After the October 2025 merger, dbt Labs publicly committed to keeping dbt Core under Apache 2.0. However, the community fears investment may shift toward dbt Fusion and dbt Cloud, leaving Core with maintenance-only patches. Some teams are evaluating alternatives due to vendor lock-in concerns.

What are the best alternatives to dbt Core?

Alternatives include dbt Cloud (managed SaaS), Dagster, Apache Airflow, Prefect, and Astronomer Astro. These platforms offer orchestration or end-to-end pipeline management that dbt Core lacks. The choice depends on whether your team needs scheduling, streaming, or a more integrated stack.

Alternatives in this category

Integrations

Snowflake BigQuery Redshift Databricks

How dbt Core compares

Direct head-to-head against 3 competitors. Picked by 7wData.

This tool

dbt Core

Pricing
dbt Core is free and open-source under Apache 2.0, with no commercial licensing fees. The core SQL transformation framework, CLI, testing, documentation, and package manager are all zero-cost. Costs come from your cloud data warehouse compute (Snowflake, BigQuery, etc.) and third-party orchestrators or hosted schedulers (Airflow, Dagster, Prefect). dbt Cloud, the managed SaaS offering, uses consumption-based pricing; historical price increases of 160–1700% have driven some teams to evaluate alternatives. dbt Core remains unlimited for self-hosted deployments.
Target
dbt Core is an open-source SQL transformation framework released under Apache 2.0, acquired in 2016 and now part of the merged Fivetran-dbt Labs company (as
Deployment
self-hosted
Strength
Free and open-source under Apache 2.0; no licensing surprises or consumption-based cost multipliers.
Watch for
Handles only transformation (the 'T' in ELT); requires separate tools for ingestion, loading, and scheduling, fragmenting stack complexity.

SQLMesh

Pricing
Open-source core free. Tobiko Cloud: platform fee plus pay-as-you-go consumption, no public tiers. Contact sales.
Target
Data engineering teams at mid-to-large orgs scaling SQL transformation, especially those hit by dbt incremental model bugs or dev-environment costs.
Deployment
Open-source self-hosted (free), plus Tobiko Cloud SaaS managed layer.
Strength
Virtual environments use views pointing to unchanged tables, so dev environments cost near zero and never drift from production state.
Watch for
Smaller hiring pool and fewer community packages than dbt. Tobiko Cloud pricing is opaque with no public tiers published.

Google Dataform

Pricing
No charge for Dataform itself. Costs billed via BigQuery compute for query execution only.
Target
Data teams running SQL transformations exclusively on BigQuery who want a free, native, low-ops tool.
Deployment
SaaS, GCP-hosted only. No self-hosted option.
Strength
Native BigQuery integration with zero licensing cost and a built-in managed scheduler inside the GCP console.
Watch for
BigQuery-only. No unit test support and no external package ecosystem, unlike dbt Core. Migrating out requires full retooling.

Coalesce

Pricing
Custom quote only. 14-day free trial. No published tiers or dollar amounts. Contact sales required.
Target
Enterprise data teams on Snowflake preferring low-code GUI over SQL-first workflows, including non-technical analysts.
Deployment
SaaS. Connects to Snowflake, Databricks, BigQuery, Redshift, and Fabric.
Strength
Automatic column propagation: adding a column upstream pushes it to all downstream models without manual YAML or macro edits.
Watch for
No native scheduler. Customers must bolt on Airflow, Snowflake Tasks, or GitHub Actions. Proprietary metadata creates migration lock-in.

User reviews

No user reviews yet. Be the first to write one.

Sources

Reporting on this tool draws on these publicly available sources.

  1. www.integrate.io — dbt Core strengths (free, open-source, SQL accessibility, testing, documentation), limitations (no scheduling, missing functionality, steep learning curve, batch-only)
  2. barc.com — User ratings (7.3/10 satisfaction), 37% missing functionality, price-to-value strength (84%), weak vendor support, ideal user profile (technical SQL teams under 2500 employees)
  3. www.fivetran.com — October 2025 merger announcement, George Fraser CEO, combined ARR approaching $600M, public commitment to keep dbt Core under Apache 2.0 license
  4. datacoves.com — Post-merger risks: license uncertainty precedent (OpenTofu, Elasticsearch), future vision favors dbt Fusion over Core, governance and vendor lock-in concerns
  5. docs.getdbt.com — dbt Core features: Fusion Engine, Semantic Layer, Mesh, testing, documentation, Jinja macros, package manager, state-aware execution
  6. mammoth.io — Criticisms: skill requirements (Git, YAML, Jinja), vendor lock-in post-acquisition, pricing shocks (dbt Cloud), limited scope (transformation only)