Business Intelligence

Lightdash

Lightdash is BI with metrics in code, in your dbt project or in Lightdash's own YAML. Flat-rate pricing, unlimited users, and way cheaper than Looker at scale.

What It Is

Lightdash is an open-source BI tool where metrics and dimensions live in YAML. You can define them in your dbt project (Lightdash's original design) or in Lightdash's own native semantic layer (added for teams without dbt). Either way, your metric definitions stay in version control, not locked inside a dashboarding UI.

Lightdash Cloud lets you create scoped AI agents directly in the UI: a marketing agent with access to campaign metrics, a sales agent tied to pipeline data, each with its own prompt and metric scope. An MCP server also lets external AI tools (Claude, Cursor, or custom agents) query the same governed semantic layer. Cloud also ships dashboards-as-code (dashboard definitions in Git) which pairs well with AI coding agents writing dashboards directly from the repo.

Why We Chose It

For teams that want version-controlled metrics plus AI analytics without Looker-tier spend, Lightdash Cloud is the right middle ground. Pricing is flat-rate for unlimited users, a fundamentally different shape from per-seat tools, which makes the economics flip in your favor once you're past ~10–15 users. A self-hosted Core edition is also available for teams that want full control.

The AI story matters too: scoped AI agents and the MCP server both query the governed semantic layer, not raw tables. When you define metrics in code, AI tools get the same definitions your dashboards use. No separate "AI data prep" needed.

Worth knowing: most of Lightdash's AI features (scoped AI agents, MCP server, dashboards-as-code, Slack assistant) are in the paid Cloud product rather than the open-source Core. The open-source Agent Skills are a separate feature: they teach coding agents how to write Lightdash YAML config, not query data.

How We Use It

Set up Lightdash Cloud or self-hosted Core deployment on GCP

Connect Lightdash to BigQuery with dbt YAML integration or Lightdash's native semantic layer

Define explores, metrics, and dimensions in dbt models or Lightdash's own YAML

Build dashboards for operational and executive reporting

Configure scheduled dashboard deliveries via Slack and email

Set up scoped AI agents per team (marketing, sales, finance), each with its own prompt and metric access

Connect Lightdash's MCP server to the client's AI tools (Claude, Cursor) for natural-language queries over governed data

Use Lightdash's dashboards-as-code workflow so coding agents can build and iterate on dashboards directly in Git

When Lightdash is the right BI tool — and when it isn't

Choose Lightdash when:

  • You want metrics defined in code, in dbt YAML or Lightdash's native semantic layer
  • You want flat-rate pricing that doesn't scale with headcount (cheaper than Looker past ~10–15 users)
  • You want to connect your own AI tools (Claude, Cursor) to governed analytics via MCP
  • Dashboards-as-code fits your team. AI coding agents can work the dashboard layer directly

Choose Looker instead when:

  • You need enterprise governance and 20+ users across departments
  • You need LookML's full semantic layer capabilities or deep BI Engine tie-in
  • Embedded analytics or multi-tenant reporting at scale is on your roadmap

Choose Metabase instead when:

  • You need broader BI (dashboards, embedding, SDK) beyond metric-first
  • Your team isn't committed to dbt-native workflows
  • You want self-host optionality with AI features included. Metabot is now open-source

Choose Steep instead when:

  • Primary audience is non-technical users
  • A metric-first UX matters more than version control
  • Budget is under €500/month for a small team

No data platform yet?

Build end-to-end from scratch — ingestion, transformations, dashboards. Audit first, then a fixed-price build.

See Implementation

Platform exists but not working?

Audit what you have, then build the improvements. Two phases — audit first, then a fixed-price build.

See Optimization

Need help with Lightdash?

Focused help on one or two layers — Dataform, Looker, BigQuery costs. Fixed-price, 2–6 weeks.

See Expert Services