dbx

dbdiagram - Database Diagram As Code

Draw ER diagrams by just writing code. Designed for developers and data analysts.

dbdocs - Database Docs As Code

Create web-based database documentation using code. Integrate seamlessly with your development workflow.

RunSQL - SQL Query Testing Tool

Quickly create mock database environments to validate your SQL queries.

dbml - Database Definition As Code

Open-source DSL language designed to define and document database schemas and structures.

Holistics.io - BI Reporting As Code

Define reusable analytics logic. Build reports and metrics using code. Version control with Git. Design and publish datasets for self-serve exploration.

Go To App 👉

Data lineage

See how data flows through your pipeline, as code

Describe your pipeline in DBML, a plain text language for data models. dbdiagram draws the flow from source to mart as you type. Built for data teams.

Draw your first pipeline

One diagram for your whole pipeline 🚀

Define data dependencies

Say how each model is built

A Dep says one table or column is built from another. Point one table at another for a quick DAG, or list the columns to say which field feeds which.

See the full syntax & use cases

// table-level: a quick DAG
Dep: raw_stripe -> stg_orders

// column-level, with a note on the transform
Dep {
  stg_orders.amount -> fct_orders.amount
  stg_orders.status -> fct_orders.state
  note: 'Keeps paid orders. Renames status to state.'
}

// inline on the target column
Table fct_orders {
  amount decimal [dep: <- stg_orders.amount]
}

Schema + flow

Schema and flow in one diagram

Foreign-key relationships (Ref) and data lineage (Dep) draw as different edge styles on the same canvas. Show them together or separately, whichever the diagram needs.

Dashed Dep edges carrying data into fct_orders, and solid Ref edges linking it to dim_customers and dim_products

Column-level lineage

Trace a column back to its source

Map dependencies column by column. Click a field in a mart and dbdiagram highlights its whole path. You see every model it passes through and the transform recorded at each step.

Group by layer

Group models into layers

Put models into color-coded groups by layer, for example sources, staging and marts (the medallion pattern). The groups come from your DBML, so a large pipeline stays readable as it grows.

See layer grouping

A 21-table pipeline grouped into four color-coded layers: raw, staging, core and marts

Design before you build

Map out a pipeline and share it for feedback before you write any SQL. Changing the design costs nothing at that stage.

Share & embed

Publish a link or embed the live diagram in Notion, Confluence or a README. Teammates open it in the browser with nothing to install.

Work in the browser or your repo

Design in the browser, or keep the .dbml in your repo and edit it in VS Code, Cursor or Windsurf with the extension. The CLI pushes and pulls, so both stay in step.

Start from your stack 🔌

Your pipeline is already defined somewhere. Generate the diagram from a dbt project or your warehouse, instead of retyping it by hand.

Coming soon

Turn a dbt project into a diagram

Point dbdiagram at a dbt project. It reads your models, sources and ref() calls and draws the lineage. Change a model and preview the diff before you ship it.

Coming soon

Connect your warehouse

Connect a database and dbdiagram reads your views and materialized views as derived models. The dependencies between them become flow edges.

FAQs

Does it show column-level lineage?

Yes. Map dependencies column by column, like stg_orders.amount -> fct_orders.amount. You can then trace one field from mart back to source, not just table to table.

How is this different from dbt docs?

dbt docs needs a warehouse connection to generate, and only shows the pipeline once it's built. dbdiagram draws the flow straight from code, so you can design and review it before the SQL exists.

Can I document the transform logic?

Yes. Attach a note to any Dep, including the SQL and how the model is materialized. The transform then sits next to the lineage it produces.

Do I need to connect my warehouse?

No. You write lineage in DBML in the editor. Connecting a database to extract views automatically is an optional path we are building.

Do I need to know DBML?

If you have used dbdiagram, you already know most of it. Lineage adds one block, Dep, next to the Table and Ref syntax you already write. If you are new, it is a small language you can pick up in minutes.

Is it free?

Yes. Writing and visualizing your lineage is free, with no limit on project size. Paid plans add advanced visualization for large diagrams, plus live connections to your warehouse or dbt project with change diffs. See pricing .

Draw your data flow, from code 😎

Write your first Dep in the editor. It takes a minute.