Metabase
Open source business intelligence that lets non-technical people ask questions of a database and analysts write SQL.
Metabase is a bi and dashboards tool for data and analytics, listed under freemium pricing and best suited to product and engineering-led companies that want self-serve dashboards on their own database.
What is Metabase?
Metabase is an open source BI tool that connects directly to your database and puts a point-and-click question builder in front of it. A non-technical user picks a table, adds filters and a summary, and gets a chart without writing SQL. An analyst opens the native SQL editor for anything the builder cannot express, saves the result as a question, and pins it to a dashboard. Dashboards support filters wired to multiple cards, drill-through, and scheduled delivery to email or Slack.
The open source edition runs as a single Java application or Docker container and is free forever. Metabase also sells a managed cloud version and paid Pro and Enterprise tiers, available either hosted or self-hosted, which add SSO through SAML and JWT, row and column level data sandboxing, white-label interactive embedding, usage analytics and official support. Models let analysts curate a clean, renamed view of a messy table so that everyone else builds on it rather than on raw columns, and metrics define reusable calculations.
Where it shines
Time to first dashboard is measured in minutes. Point the container at Postgres, MySQL, Snowflake, BigQuery, Redshift, Databricks, MongoDB, ClickHouse or any of the other supported databases, and the automatic exploration of each table gives you charts before you have configured anything. Startups routinely run Metabase against a read replica of the production database on day one.
The question builder is genuinely usable by people who do not know SQL. Because it exposes the database schema through a guided interface rather than a drag-and-drop canvas, support and operations staff can answer their own questions about customers or orders and stop filing tickets for the data team.
Embedding is a strong second use case. Static embeds put a filtered chart in your own product cheaply, and interactive embedding on the paid tiers hands customers a full, sandboxed query experience under your branding. Compared with the embedded analytics products aimed at enterprises, the price and setup effort are small.
Self-hosting on your own infrastructure with an open licence keeps data inside your network and avoids per-seat pricing. The free tier has no user limit, which is rare in this category.
Where it falls short
Metabase is a tool for querying databases, not a data platform. There are no SaaS connectors for Google Analytics, Salesforce or ad platforms, so data has to be landed in a warehouse first by a tool such as Fivetran. Teams expecting a connector marketplace like Domo will be disappointed.
Visualisation is functional rather than beautiful. Chart types cover the essentials, formatting options are limited, and there is nothing like the mark-level control that Tableau offers. Large tables with many dashboard cards can feel slow because every card runs its own query against the source with limited caching.
Modelling depth is modest. Models and metrics help, but there is no full semantic layer with joins defined once and reused everywhere, and complex multi-fact analysis still ends up in hand-written SQL. Row-level sandboxing, SSO and audit logs are paid features, so a company that needs them loses the zero-cost advantage. Pricing for the paid tiers jumps in steps rather than scaling smoothly with seats, which surprises mid-sized teams.
Conclusion
Metabase is the best open source BI tool for companies whose data already sits in a database or warehouse and who want dashboards without a procurement process. Engineering-led startups and product teams get the most from it, and the embedding story makes it a practical choice for customer-facing analytics. Buy something else if you need connectors to SaaS sources, a deep semantic model, or presentation-grade visuals.