BI & Dashboards

This is the "I have a Postgres / warehouse and want a Looker-shaped tool" page. For product / web analytics that ship their own collection layer (PostHog, Plausible, Umami) see Analytics. For raw chart libraries you build your own UI on top of, see Charts.

Open-source BI platforms

• ★ Metabase — the dominant OSS BI tool. SQL editor, no-code question builder, dashboards, alerts, embedding. AGPLv3 (self-host free) plus paid Cloud / Enterprise. Default pick if you want Looker-ish for free.
• ★ Apache Superset — Airbnb-origin, Apache-licensed; enormous chart library, SQL Lab, Jinja-templated dashboards. Heavier ops than Metabase but more flexible. Preset.io is the hosted commercial version.
• Redash — query-centric BI; lighter than Superset; BSD-3. Less actively maintained since the Databricks acquisition but still widely deployed.
• Lightdash — dbt-native BI; metrics live in your dbt project (meta: blocks become explore-able dimensions/measures). MIT + paid Cloud. The clearest "BI built on dbt" pick in 2026.
• Evidence — code-based BI: Markdown + SQL → static dashboards. Git-versioned, CI-deployable. MIT; great for data teams that want PR-reviewed reports.
• Rill Data — DuckDB-powered, dashboard-first; instant slice-and-dice on local Parquet / S3. Apache 2.0; hosted Rill Cloud.
• Briefer — newer; notebooks + dashboards + scheduled reports in one app. Open source (AGPL) + hosted.

Grafana for non-ops dashboards

• ★ Grafana — yes, it's an ops tool (observability, ops/log-aggregation), but it's also fine for business dashboards on Postgres / ClickHouse / BigQuery / Snowflake. AGPLv3 since 2021. Worth considering when the rest of your stack is already Grafana-shaped.

Hosted free tiers worth knowing

• Metabase Cloud Starter — small free trial; usually self-host wins on cost.
• Preset.io (managed Superset) — free tier limited; pay per workspace.
• Looker Studio (formerly Data Studio) — free, Google-flavored, light; fine for marketing dashboards on GA4 / BigQuery.
• Mode, Hex, Deepnote — notebook-driven analytics with free tiers; see Notebooks.

Code-based / "BI as code"

• ★ Evidence — Markdown + SQL; the leading code-first BI tool.
• Lightdash — metrics defined in dbt YAML.
• Vizro (McKinsey OSS) — Python config → Plotly Dash apps; light declarative BI.
• Observable Framework — Markdown + JS data apps; static-export dashboards; see Notebooks.

License / pricing watch-outs

• Metabase is AGPLv3 — fine to self-host internally; review legal if you embed it externally without commercial license.
• Superset is Apache 2.0 — fully permissive.
• Lightdash is MIT for the OSS edition; embedded analytics features are paid.
• Redash is BSD-3; community is quieter than it was.
• Grafana is AGPLv3 since 2021.

Pick this if…

• Default OSS BI, easy ops: Metabase.
• You want depth and don't mind ops: Apache Superset.
• Your team lives in dbt: Lightdash.
• You want Git-reviewable, code-only dashboards: Evidence.
• Marketing-y dashboards on Google data: Looker Studio.
• Already running Grafana: Grafana with a Postgres / ClickHouse data source.
• Sub-second slice-and-dice on Parquet / S3: Rill.