Metabase vs Apache Superset

Quick verdict

Metabase is a BI tool built by Metabase, Inc. The GitHub repository contains both the open source edition, released under the AGPL, and the commercial editions under the Metabase Commercial Software License. The latest release on GitHub when checked was 63.19 (1 October 2026), with a 64.0 beta in testing. You can self-host it or buy Metabase Cloud, on Starter, Pro and Enterprise plans.

Apache Superset is an Apache Software Foundation project under the Apache License 2.0. It describes itself as a data exploration and visualisation platform with a no-code chart builder, a web SQL editor (SQL Lab) and a lightweight semantic layer of datasets, metrics and calculated columns. Version 6.1.0 was released on 13 May 2026 and requires Python 3.10 or later. Preset is a separate company that sells a managed Superset service; there is no separate paid edition of Superset itself.

Side by side

Key differences

Licences and editions

Metabase is "open core". The open source edition is AGPL: you can run it internally for free, but if you modify it and make it available to others over a network, the AGPL requires you to offer the modified source. Features such as SSO, row and column-level permissions, SSO-based embedding and advanced theming are commercial and need a paid plan, whether you self-host or use Metabase Cloud.

Apache Superset has one edition under the permissive Apache License 2.0, so there is no feature gating inside the project and no copyleft obligation if you modify or embed it. Commercial services come from third parties, most visibly Preset, which adds hosting, multiple workspaces, support and, on its Enterprise plan, features such as SSO/SCIM and audit logs on top of Superset.

SQL editors: native queries versus SQL Lab

Metabase's native query editor is built to turn a query into a reusable question. You add filter widgets with variables and field filters, save reusable SQL as snippets, and refer to models or saved questions as if they were tables. Metabase documents that it runs single read statements only: no multi-statement scripts, stored procedures or DDL.

Superset's SQL Lab is closer to a SQL IDE in the browser. Superset's documentation describes it as the place for "cleaning, joining, and preparing data" for the Explore workflow. A SQL Lab query can be saved as a virtual dataset, which then feeds charts and dashboards like a table.

-- A query you might save as a virtual dataset in Superset
-- or as a SQL model in Metabase (PostgreSQL)
SELECT c.region,
       date_trunc('month', o.ordered_at) AS month,
       sum(o.amount) AS revenue
FROM orders o
JOIN customers c ON c.id = o.customer_id
GROUP BY c.region, date_trunc('month', o.ordered_at);

In Metabase the same query can become a model, a curated starting point for business users who then filter and summarise it in the graphical builder. Metabase notes that when a model is written in SQL you fill in column metadata manually, whereas a builder-based model gets it automatically. In our view, Metabase suits mixed audiences, while Superset suits teams where most users are comfortable writing SQL.

Database connectivity

Metabase ships official drivers for Athena, BigQuery, ClickHouse, Databricks, Druid, MongoDB, MariaDB, MySQL, Oracle, PostgreSQL, Presto, Redshift, Snowflake, SparkSQL, SQL Server, SQLite, Starburst and Vertica, with community drivers for others. Official support for these drivers comes with the Pro and Enterprise plans.

Superset connects through SQLAlchemy: each engine needs a SQLAlchemy dialect and a Python DB-API driver. Its documentation lists 81 databases, but states that Superset "does not ship bundled with connectivity to databases", so in a Docker install you add driver packages (for example psycopg2-binary) to requirements-local.txt or build a custom image.

Self-hosting requirements

Metabase is a single Java application. The docs recommend Eclipse Temurin JRE 25 for the JAR, or running the Docker image, and advise replacing the default H2 application database with PostgreSQL or MySQL in production.

Superset offers Docker Compose (for testing and development), Kubernetes (which its docs call "the best-practice way to deploy a production instance") and PyPI installs. A production deployment needs a PostgreSQL or MySQL metadata database, Redis for caching, Celery workers and a beat scheduler, and a custom image that adds your database drivers and a headless browser for Alerts and Reports. In our view this is a real operational cost for a small team, and is the main reason people choose Preset.

Embedding

Metabase supports public links, guest (signed) embeds and full-app embedding on every plan, including open source. Modular embedding with SSO, the React-based Embedded analytics SDK features, row-level data segregation and advanced theming need Pro or Enterprise.

Superset provides the @superset-ui/embedded-sdk package. You enable the EMBEDDED_SUPERSET feature flag, your backend requests a short-lived guest token from Superset's API with a service account, and you can attach row-level security rules to that token. This is part of the open source project. Preset sells embedded dashboards as an add-on.

Pricing and licensing

Metabase. The open source edition is free with unlimited users. Listed on Metabase's pricing page in October 2026 (USD): Starter USD 100 per month with 5 users included, then USD 6 per user per month; Pro USD 575 per month with 10 users included, then USD 12 per user per month; Enterprise custom, from USD 20,000 per year. Yearly billing is discounted, and Starter and Pro have a 14-day trial. Both cloud and self-hosted deployments are available on paid plans.

Apache Superset and Preset. Superset is free. Listed on Preset's pricing page in October 2026 (USD): Starter is free for up to 5 users; Professional is USD 20 per user per month billed annually or USD 25 billed monthly, with a 14-day trial; Enterprise is quoted. Embedded dashboards are an add-on starting at USD 500 per month for 50 viewer licences. Self-hosting Superset costs nothing in licences, but you pay for the infrastructure and the people who run it.

Pricing checked on the vendors' official pages on 7 October 2026. Prices change; confirm before buying.

Where each one leads

Limitations

When to choose each

When neither is right

• Your metrics are defined in dbt and you want BI on top of that project: look at Lightdash in the best open source BI tools.
• You want reports written as code in SQL and Markdown: Evidence, also in the open source BI guide.
• You need an enterprise semantic model and Microsoft integration: see Power BI vs Metabase and Power BI vs Tableau.
• You need a governed semantic layer defined in code: see Power BI vs Looker and Tableau vs Looker.

Final recommendation

Frequently asked questions

Sources

• Metabase pricing
• Metabase on GitHub (licences)
• Metabase releases
• Metabase: connecting to databases
• Metabase: SQL editor
• Metabase: models
• Metabase: embedding overview
• Apache Superset on GitHub
• Apache Superset on PyPI
• Superset: connecting to databases
• Superset: installation methods
• Superset: embedding
• Preset pricing

Checked October 2026.

How we research comparisons: our editorial method.