Lightdash

What is Lightdash?

Lightdash is an open source business intelligence tool that turns a dbt project into a self-serve analytics layer. Instead of redefining metrics in a separate BI app, teams write dimensions and metrics as YAML alongside their dbt models, and Lightdash exposes them as explorable dashboards, charts, and AI agents — with everything governed in one place and managed through Git.

What is Lightdash best for?

Lightdash is best for data teams already invested in dbt who want their metric definitions to live in version control rather than scattered across a BI tool’s UI. It suits analytics engineers who treat BI as code — reviewing metric changes in pull requests, shipping through CI, and keeping a single governed source of truth. If you don’t use dbt, it’s a poor fit.

What can Lightdash do?

• Define metrics and dimensions as code in YAML next to your dbt models, so they live in Git and change through pull requests.
• Auto-generate explorable dimensions from dbt models, giving business users point-and-click self-serve exploration without writing SQL.
• Build dashboards and charts that reference the shared metric layer, keeping numbers consistent across every report.
• Answer questions with AI agents that query the governed context layer in natural language while respecting permissions.
• Connect to many warehouses — BigQuery, Snowflake, Redshift, Databricks, PostgreSQL, Trino, and ClickHouse among them.
• Embed analytics into your own apps via SDKs with row-level security.
• Schedule reports and deliver them to Slack or email, plus a CLI and MCP support for shipping analytics like software.

Where does Lightdash fall short?

• It is tied to dbt. The metric-as-code workflow assumes a dbt project and dbt fluency; a newer standalone YAML mode exists, but teams without dbt expertise get little of the benefit and are usually better off with Metabase.
• Visualization and formatting are less mature than long-established tools. Chart types, complex calculated fields, and pixel-perfect report layouts are more limited than Tableau, Power BI, or even Metabase.
• The community is smaller. Fewer tutorials, third-party integrations, and Stack Overflow answers exist than for older BI tools, so you lean more on the official docs and community Slack when you get stuck.

Is Lightdash free?

Partly. The open source edition is MIT-licensed and free to self-host, with the core BI features — metrics, exploration, dashboards, and warehouse connections. Lightdash also sells a managed Cloud Pro plan (listed at $3,000/month with unlimited users, native dbt integration, scheduled reports, and AI agents) and a custom-priced Enterprise tier adding SSO/SAML, SCIM, custom roles, and SOC 2 compliance. Some enterprise features (code under packages/backend/src/ee) carry a separate commercial license rather than MIT.

What does Lightdash replace?

Lightdash stands in for governed, enterprise BI platforms — most directly Looker, whose LookML “metrics as code” model it echoes, and to a lesser extent Tableau and Power BI for teams that want dashboards and self-serve analytics without per-seat licensing. Among open source options it’s closest to Metabase, though the two target different audiences: Metabase for out-of-the-box non-technical self-serve, Lightdash for dbt-native governance.

FAQ

Is Lightdash open source? Yes. The core is MIT-licensed and the source is on GitHub. Some enterprise-edition features are held under a separate commercial license, a common open-core arrangement.

Can I self-host Lightdash for free? Yes. The open source edition self-hosts free via Docker or Kubernetes, backed by a PostgreSQL database, with community Slack support. Paid Cloud and Enterprise tiers are optional.

Is Lightdash a good Looker alternative? For dbt teams, yes — it delivers a similar governed, metrics-as-code model without Looker’s licensing cost. It’s less polished on advanced visualization and has a smaller ecosystem, so weigh that against the price savings.

What do I need to run Lightdash? A dbt project (or standalone Lightdash YAML), a supported data warehouse, and a host for the app plus a PostgreSQL metadata database. Self-hosting runs on Docker or Kubernetes.