Lightdash
Open-source BI tool that builds directly on dbt models and metrics, exposing a REST API for programmatic chart and dashboard access — enabling agents to query curated business metrics without raw SQL.
Score Breakdown
⚙ Agent Friendliness
🔒 Security
Row-level security inherited from dbt model definitions. API tokens are not scoped to specific projects — a leaked token grants access to all projects the user can see.
⚡ Reliability
Best When
Your team already uses dbt and wants agents to query governed, version-controlled business metrics through a REST API rather than raw warehouse SQL.
Avoid When
You don't have a dbt project or need to define metrics outside of dbt's YAML-based metric layer.
Use Cases
- • Agent fetches current KPI values (revenue, DAU, churn) via Lightdash REST API using pre-defined dbt metrics without writing SQL
- • Automate dashboard screenshot or data export tasks by calling Lightdash's scheduler and download endpoints from an agent workflow
- • Agent discovers available metrics and dimensions in a project using the catalog API to dynamically build exploratory queries
- • Embed Lightdash chart iframes in agent-generated reports by fetching share URLs programmatically via the API
- • Trigger Lightdash project refreshes after dbt model runs complete, keeping agent-facing metrics current without manual intervention
Not For
- • Ad-hoc SQL exploration without a dbt project — Lightdash requires dbt models as its semantic layer foundation
- • Real-time streaming analytics dashboards requiring sub-second refresh — Lightdash queries run synchronously against the underlying data warehouse
- • Non-technical end users who need a drag-and-drop BI tool without any dbt knowledge for metric definition
Interface
Authentication
Personal API tokens generated in user settings. Service accounts available for programmatic access. Token passed as Authorization: ApiKey <token> header.
Pricing
MIT-licensed open source with optional managed Lightdash Cloud. Self-hosting is production-ready with Docker Compose or Kubernetes helm chart.
Agent Metadata
Known Gotchas
- ⚠ Query results are paginated but the max page size is not documented — agents requesting large result sets may need multiple round trips without clear guidance
- ⚠ Metric queries require knowing the explore name and field names exactly — there is no fuzzy search; agents must call the catalog API first to discover valid field identifiers
- ⚠ Dashboard and chart UUIDs are not stable across project refreshes when items are deleted and recreated — agents storing UUIDs should validate them before use
- ⚠ dbt metric definitions must be compiled and synced before they appear in Lightdash — agents triggering metrics queries immediately after dbt run may see stale catalog
- ⚠ API tokens are user-scoped and inherit that user's row-level security filters — agent queries may return different data than expected if the service account user has restricted access
Alternatives
Full Evaluation Report
Detailed scoring breakdown, competitive positioning, security analysis, and improvement recommendations for Lightdash.
Scores are editorial opinions as of 2026-03-06.