Lightdash — Open Source BI Platform for dbt Users

Introduction

Lightdash is an open-source BI tool built specifically for teams using dbt. It reads your dbt models, metrics, and descriptions so analysts can explore data and build dashboards without re-defining business logic in a separate tool.

What Lightdash Does

• Syncs with your dbt project to import models, dimensions, and metrics automatically
• Provides a point-and-click explore interface for ad-hoc queries and data exploration
• Enables self-serve dashboards that business users can build and share
• Generates SQL under the hood while keeping the dbt semantic layer as the source of truth
• Supports scheduling, alerts, and Slack or email delivery of reports

Architecture Overview

Lightdash connects to your data warehouse (BigQuery, Snowflake, Redshift, PostgreSQL, Databricks, or Trino) and reads dbt manifest artifacts to understand your data model. Queries are compiled to SQL and executed directly against the warehouse. The web application is a Node.js backend with a React frontend, and state is stored in a PostgreSQL metadata database.

Self-Hosting & Configuration

• Deploy with Docker Compose or Helm charts for Kubernetes
• Connect to your data warehouse with credentials and a dbt project repository
• Run dbt compile to generate the manifest that Lightdash reads
• Configure SSO via Google, Okta, or generic OIDC for team authentication
• Set up scheduled deliveries and alerts via Slack or email integrations

Key Features

• Native dbt integration keeps metrics defined in one place with no logic duplication
• Interactive explore mode for building queries by clicking dimensions and metrics
• Shareable dashboards with filters, cross-filtering, and drill-downs
• Role-based access control and row-level security for data governance
• CLI tool for developers to preview changes and deploy from CI pipelines

Comparison with Similar Tools

• Metabase — General-purpose BI tool with its own model layer; Lightdash uses dbt as the single source of truth
• Apache Superset — Powerful visualization platform but requires separate metric definitions outside dbt
• Looker — Commercial BI with LookML modeling; Lightdash achieves similar goals with dbt and open source
• Evidence — Code-first BI using SQL and Markdown; Lightdash offers a GUI-based explore experience
• Redash — SQL-focused dashboarding tool without native dbt integration

FAQ

Q: Do I need dbt to use Lightdash? A: Yes. Lightdash is built around the dbt semantic layer and requires a dbt project.

Q: Which data warehouses are supported? A: BigQuery, Snowflake, Redshift, PostgreSQL, Databricks, Trino, and DuckDB.

Q: Can non-technical users build their own dashboards? A: Yes. The explore interface is designed for self-serve use without writing SQL.

Q: Is there a managed cloud version? A: Yes. Lightdash offers a hosted cloud option alongside the self-hosted open-source edition.

Sources

• https://github.com/lightdash/lightdash
• https://docs.lightdash.com