dbt Core is completely free and open-source. dbt Cloud offers a free developer tier and paid team/enterprise plans starting at $50/month per seat.

What databases does dbt support?

dbt supports all major cloud warehouses including Snowflake, BigQuery, Redshift, Databricks, DuckDB, and PostgreSQL via adapters.

Do I need to know Python to use dbt?

No. dbt is SQL-first. You write models in SQL with Jinja templating. Python models are supported in dbt Core 1.3+ but are optional.

What's the difference between dbt Core and dbt Cloud?

dbt Core is the open-source CLI you run locally or in CI. dbt Cloud adds a hosted IDE, job scheduler, observability, and team collaboration features.

Where does dbt fit in the modern data stack?

dbt sits in the transformation layer — after data ingestion tools like Fivetran or Airbyte load raw data into your warehouse, dbt transforms it into analytics-ready models.

What is dbt?

The complete guide for data engineers — what it does, how it works, and when to use it.

10 min read · Updated March 2026

Start Learning dbt →Build a dbt Pipeline →

Quick Answer

dbt (data build tool) is an open-source SQL transformation framework used by data engineers and analytics engineers to build modular, tested, and version-controlled data models inside a data warehouse. It does not move data — it transforms data that's already been loaded, following the ELT pattern. dbt is maintained by dbt Labs and is the industry standard for analytics engineering.

What is dbt?

dbt is an open-source command-line tool (and optional cloud platform) that lets data teams write SQL transformations as version-controlled code. Instead of writing ad-hoc SQL queries or building fragile stored procedures, you define models — plain SQL files that dbt compiles and runs against your data warehouse.

dbt is SQL-first. Every transformation is a .sql file with optional Jinja templating. dbt generates the DDL/DML, manages dependencies between models, and materializes results as tables or views in your warehouse.

Critically, dbt follows the ELT pattern (Extract, Load, Transform), not ETL. It does not move data — tools like Fivetran, Airbyte, or Stitch handle ingestion. dbt only transforms data that has already landed in your warehouse.

dbt CoreFree & Open-source

The open-source CLI. Run it locally, in Docker, or in CI/CD. Full feature set with no restrictions. The foundation that dbt Cloud is built on.

dbt Cloud$50+/mo

Managed SaaS by dbt Labs. Adds a browser IDE, job scheduler, CI/CD integrations, model observability, and team collaboration. Free developer tier available.

Why dbt matters:

Version control for SQL — every model lives in Git, reviewed like code
Built-in testing — assert uniqueness, nulls, referential integrity on any column
Auto-generated documentation — lineage graphs, column descriptions, and data catalog
Dependency management — dbt resolves model dependencies automatically via ref()
Team collaboration — SQL organized in folders, reviewed in PRs, deployed in CI

Why dbt matters

Before dbt

Messy SQL scripts with no structure
No version control for transformations
Hard to test data quality
Duplicated logic across reports
Tribal knowledge — no documentation

With dbt

Modular SQL organized in folders
Git-based version control and code reviews
Automated data quality tests
Reusable macros and packages
Auto-generated docs and lineage graphs

What you can do with dbt

Build clean staging, intermediate, and mart models

Define business metrics in a semantic layer

Power BI dashboards with tested, reliable tables

Implement data contracts and schema validation

Create ML feature tables for training pipelines

Document your entire data warehouse automatically

How dbt works

The ELT flow with dbt looks like this: raw data is loaded into your warehouse first (via Fivetran, Airbyte, Stitch, or custom pipelines), then dbt transforms it in-warehouse into clean, analytics-ready tables.

Source Systems

Postgres, APIs, S3

›

Ingestion

Fivetran / Airbyte

›

Raw Layer

Warehouse tables

›

dbt Transform

Models + Tests

›

Analytics Layer

BI / AI tools

Three core concepts:

1. Models

Each .sql file in your models/ directory is a model. dbt runs them in dependency order and materializes results as tables or views.

2. Tests

Define data quality assertions in YAML. Built-in tests include unique, not_null, accepted_values, and relationships. Custom tests can be written in SQL.

3. Documentation

dbt auto-generates a documentation site from your YAML descriptions, with lineage DAGs showing how every model relates to its sources and downstream consumers.

models/staging/stg_orders.sql

-- models/staging/stg_orders.sql
with source as (
    select * from {{ source('ecommerce', 'raw_orders') }}
),
renamed as (
    select
        id          as order_id,
        user_id,
        status,
        created_at
    from source
)
select * from renamed

models/staging/schema.yml

# schema.yml
models:
  - name: stg_orders
    columns:
      - name: order_id
        tests:
          - unique
          - not_null

Common mistakes with dbt

Putting business logic in sources instead of models
Not using ref() — hardcoding table names breaks lineage
Skipping tests on critical models
Monolithic models instead of modular staging → intermediate → mart layers
Using dbt as an ingestion tool (it only transforms, use Airbyte/Fivetran for ingestion)

Who should learn dbt?

Junior Data Engineers

Learn basic transformations, understand data modeling, build your first production pipeline.

Senior Data Engineers

Design scalable dbt projects, implement testing frameworks, govern data contracts.

Staff / Lead Engineers

Define the semantic layer, standardize models across the org, lead dbt adoption.

dbt vs other tools

dbt vs Apache Spark

dbt

SQL-based, warehouse-native
Analytics transformation
Free (Core)
Runs inside your warehouse

Apache Spark

Distributed compute
Python / Scala
Large-scale processing
Complex infrastructure

Verdict: Use dbt for analytics transformation inside your warehouse. Use Spark for large-scale data processing that needs distributed compute.

dbt vs Apache Airflow

dbt

Transformation logic
What the data becomes
SQL-first models
Data quality testing

Apache Airflow

Orchestration
When and how pipelines run
DAG scheduling
Task dependency management

Verdict: They are complementary. Airflow triggers dbt jobs. dbt does the transformation. Most production stacks use both.

dbt vs raw SQL scripts

Raw SQL scripts

Manual execution
Hard to test
No lineage
No version control
Breaks silently

dbt

Modular and reusable
Tested automatically
Versioned in Git
Documented with lineage
CI/CD ready

Verdict: dbt is what happens when you apply software engineering practices to SQL.

dbt Core vs dbt Cloud vs alternatives

Tool	Type	SQL-first	Testing	Lineage	Cost
dbt Core	OSS CLI	✓	✓	✓	Free
dbt Cloud	Managed SaaS	✓	✓	✓	$50+/mo
SQLMesh	OSS + Cloud	✓	✓	✓	Free/Paid
Dataform	Google Cloud	✓	✓	✓	GCP pricing

Frequently asked questions

Is dbt free?: dbt Core is completely free and open-source. dbt Cloud offers a free developer tier and paid team/enterprise plans starting at $50/month per seat.
What databases does dbt support?: dbt supports all major cloud warehouses including Snowflake, BigQuery, Redshift, Databricks, DuckDB, and PostgreSQL via adapters.
Do I need to know Python to use dbt?: No. dbt is SQL-first. You write models in SQL with Jinja templating. Python models are supported in dbt Core 1.3+ but are optional.
What's the difference between dbt Core and dbt Cloud?: dbt Core is the open-source CLI you run locally or in CI. dbt Cloud adds a hosted IDE, job scheduler, observability, and team collaboration features.
Where does dbt fit in the modern data stack?: dbt sits in the transformation layer — after data ingestion tools like Fivetran or Airbyte load raw data into your warehouse, dbt transforms it into analytics-ready models.