Lightdash

Lightdash makes dbt models self-serve — teams define metrics in dbt, and Lightdash exposes them as explorable dimensions and measures in a BI interface and REST API. For agents: query any dbt-defined metric via the REST API without writing SQL; Lightdash handles the SQL generation and caching. Self-hostable or managed. Strong fit for data teams already using dbt who want to expose metrics to non-SQL users or AI agents. Confidence is docs-derived.

REST API at app.lightdash.com/api/v1 (Cloud) or self-hosted. Resources: projects, explores, charts, dashboards, saved-charts. GET /projects/{id}/explores lists available explores (from dbt models). POST /projects/{id}/explores/{explore}/runQuery runs a query ({dimensions: [...], metrics: [...], filters: {...}}). GET /projects/{id}/charts lists saved charts. Response: JSON with rows array. API enables programmatic metric querying without SQL knowledge.

Personal access token auth: Authorization: ApiKey {token} header. Tokens from Lightdash account settings → Personal Access Tokens. HTTPS enforced. No OAuth2. Workspace-scoped queries. Rate limiting not prominently documented.

Query errors: 400 with detail (unknown dimension, invalid filter). Database query timeouts surface as errors. Explore query performance depends on underlying dbt model and warehouse query optimization. Lightdash Cloud uptime at status.lightdash.com. Self-hosted: operator responsible for reliability.

docs.lightdash.com covers setup (dbt connection), REST API reference, self-hosting guide, and dashboard creation. Getting started: connect Lightdash to a dbt project and warehouse, explore first model in under 30 minutes. Open-source MIT (GitHub: lightdash/lightdash). Community via Lightdash Slack.