Tableau vs Apache Superset: SQL, Hosting, Licences (2026)

Quick verdict

Short answer

Choose Tableau if analysts explore data visually and iteratively, need extracts to work offline from slow or remote sources, want a desktop app on Windows or macOS, and you can budget for per-role licences. Choose Apache Superset if your analysts start from a SQL query rather than a canvas, your data is already in a SQL database or warehouse fast enough to query live, and you prefer infrastructure costs to licence costs. Superset is a dashboarding layer on top of SQL; Tableau is an exploration environment that hides the SQL.

How we know: This comparison is research-based: licensing models, platforms, connectivity, SQL access, calculation features, self-hosting and embedding were checked against Tableau's help documentation and product pages, the Apache Superset documentation and GitHub repository, and Preset's pricing page in October 2026. Tableau's pricing page could not be read automatically, so no Tableau prices are quoted. We have not deployed either product and make no performance claims.

Tableau is a visual analytics platform owned by Salesforce. Authors drag fields onto a canvas in Tableau Desktop (Windows and macOS) or in web authoring, and Tableau generates the queries. Data can stay in the source through a live connection or be copied into a Tableau extract. Work is published to Tableau Cloud (SaaS) or self-hosted Tableau Server (Windows or Linux), licensed by role: Creator, Explorer and Viewer. Tableau Public is a free option for data you are willing to publish openly.

Apache Superset is an Apache Software Foundation project under the Apache License 2.0, current at version 6.1.0 (13 May 2026). It runs in the browser and pairs a SQL IDE, SQL Lab, with a no-code chart builder over datasets: physical tables, or saved SQL queries known as virtual datasets, with SQL metrics and calculated columns. It has no extract engine; every chart queries the source database. Preset is a separate company selling managed Superset.

Side by side

Aspect	Tableau	Apache Superset
Licence	Proprietary; role-based licences (Creator, Explorer, Viewer), annual contracts	Apache License 2.0; no paid edition of the project
Free option	Tableau Public (published data is public); trials of paid products	The whole project; Preset Starter is free for up to 5 users
Authoring	Tableau Desktop (Windows, macOS) and web authoring	Browser only, on any operating system
Primary workflow	Drag fields onto a canvas; Tableau writes the queries	Write SQL in SQL Lab or pick a table, save a dataset, then build charts
SQL access	Custom SQL on most database connections: a single SELECT, wrapped in an outer query	SQL Lab with schema browser and dialect-aware formatting; optional Jinja templating in queries
Calculations	Calculated fields, table calculations and level of detail (LOD) expressions	Metrics and calculated columns written as SQL expressions in the database's dialect
Live or copied data	Live connections or extracts	Live queries only, with result caching through Redis
Connectivity	Named connectors plus generic Other Databases (JDBC) and (ODBC); Tableau Bridge for private networks with Tableau Cloud	SQL databases through SQLAlchemy dialects and Python drivers; docs cover 81 databases
Self-hosting	Tableau Server on Windows or Linux, licensed by role	Free to self-host: metadata database, Redis, Celery workers; Kubernetes recommended for production
Main trade-off	Rich visual exploration and extracts, but licence cost for every author and viewer	No licence fees and SQL-native, but fewer exploration features and more to operate

Key differences

Exploration model: canvas first versus SQL first

Tableau is built for iterative visual exploration: an analyst drags dimensions and measures onto rows, columns and marks, and Tableau generates the queries against a live connection or an extract. Its own calculation language covers what the canvas cannot, including level of detail (LOD) expressions that fix the grain of an aggregation independently of the view. This example from Tableau's help returns total sales per region whatever else is on the canvas:

-- Tableau calculated field (FIXED level of detail expression)
{FIXED [Region] : SUM([Sales])}

Superset starts from SQL. The same per-region total is a virtual dataset written in the database's own dialect (shown here as standard SQL that runs on PostgreSQL), saved from SQL Lab and then charted:

-- Superset virtual dataset (PostgreSQL)
SELECT o.region,
       o.order_id,
       o.sales,
       sum(o.sales) OVER (PARTITION BY o.region) AS region_sales
FROM orders o;

In our assessment, Tableau rewards analysts who explore by trial and error and do not want to write SQL for each question, while Superset rewards analysts who are faster in SQL than on a canvas and want the logic to be readable SQL. If window functions like the one above are new to you, see our SQL functions reference.

SQL access in detail

Tableau supports custom SQL on most database connections, with documented rules: it must be a single SELECT, Tableau wraps it in an outer query to apply filters and grouping (which Tableau notes can affect performance), and parameters can replace literal values but not table names. SQL is an input to Tableau, not the place where most work happens.

Superset's SQL Lab is a browser SQL IDE with a collapsible schema tree, sample data, find and replace, code folding and a Format SQL button that applies the dialect of the selected database. Results become charts only after being saved as a dataset. With the ENABLE_TEMPLATE_PROCESSING flag (off by default), SQL can use Jinja macros such as filter_values() so dashboard filters reach the query; Superset warns that templates run on the server, so only trusted users should edit datasets when it is enabled.

Data connectivity: extracts and private networks

Tableau offers named connectors for common databases and warehouses, plus generic Other Databases (JDBC) and (ODBC) connectors. Tableau states it gives no guarantee that a particular JDBC driver will work, and some drivers support only extract creation. Extracts copy data into Tableau so dashboards do not hit the source on every click. For Tableau Cloud, Tableau Bridge (Windows, or Linux in containers) reaches data inside a private network for live queries and scheduled extract refreshes.

Superset connects only to SQL databases, through a SQLAlchemy dialect and DB-API driver for each; its docs cover 81 databases, from PostgreSQL, SQL Server and Snowflake to Trino and BigQuery. It does not bundle drivers, so you add them to your image. Because it has no extract engine, Superset depends on the source being fast enough to query interactively, which in our view makes it best suited to a warehouse or a well-indexed reporting database; see data warehouse vs database.

Self-hosting and operations

Tableau Server is a supported commercial server for Windows or Linux, installed and upgraded by your team, and licensed with the same Creator, Explorer and Viewer roles as Tableau Cloud. You get vendor support, but self-hosting does not remove licence costs. Superset costs nothing to license when self-hosted, but its docs recommend Kubernetes for production, with a PostgreSQL or MySQL metadata database, Redis, Celery workers and a beat scheduler, and a headless browser for Alerts and Reports. Support comes from the community or from Preset if you use its managed service.

Embedding and audience licensing

Tableau's Embedding API v3 embeds views and web authoring into web pages using a <tableau-viz> web component, for Tableau Cloud, Server and Public. Embedded viewers need a licence (for example the Viewer role) or a usage-based arrangement documented in Tableau's licensing pages. Superset's embedded SDK, part of the open source project, uses short-lived guest tokens requested by your backend, optionally with row-level security rules, behind the EMBEDDED_SUPERSET feature flag; there is no per-viewer fee. Preset sells embedded dashboards as a priced add-on on its managed service.

Pricing and licensing

Tableau. Tableau publishes per-role list prices (Creator, Explorer, Viewer, billed annually) on its pricing page, which we could not load to confirm current figures, so none are quoted here. What Tableau documents: every deployment needs at least one Creator, roles are sold in Standard and Enterprise editions, Tableau Cloud and Tableau Server require annual contracts, and usage-based and capacity-based options are available through sales. Tableau Public is free for public data. Check Tableau's page for current prices in your currency.

Apache Superset and Preset. Superset is free under the Apache License 2.0; self-hosting costs are infrastructure and staff time. Listed on Preset's pricing page in October 2026 (USD): Starter free for up to 5 users; Professional USD 20 per user per month billed annually (USD 25 monthly), with a 14-day trial; Enterprise quoted; embedded dashboards an add-on from USD 500 per month for 50 viewer licences.

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

Where each one leads

Tableau strengths

Drag-and-drop visual exploration with LOD expressions for controlling aggregation grain
Extracts as well as live connections, so dashboards need not query the source on every click
Tableau Desktop on Windows and macOS, plus web authoring
Vendor-supported self-hosting on Windows or Linux, and Tableau Bridge for private data with Tableau Cloud
Free Tableau Public for openly published work

Apache Superset strengths

Apache 2.0 licence with no paid edition and no per-viewer fees
SQL Lab and SQL-defined virtual datasets keep logic in readable SQL
Connects to any SQL database with a SQLAlchemy dialect
Embedded SDK with guest tokens and row-level security in the open source project
Managed option from Preset with a free Starter tier

Limitations

Tableau limitations

Every author and viewer needs a paid role or a usage-based arrangement, on annual contracts
Custom SQL must be a single SELECT and is wrapped in an outer query
Logic spread across workbook calculations can be hard to review as code (editorial)
Self-hosting Tableau Server still requires licences and server administration

Apache Superset limitations

No extract engine: the source database must handle interactive query load
SQL databases only; each driver installed separately
Production deployment needs Redis, Celery workers, a metadata database and ideally Kubernetes
Less suited to analysts who do not write SQL (editorial)

When to choose each

Choose Tableau if

Analysts explore visually and iteratively rather than starting from SQL
You need extracts because sources are slow, remote or rate-limited
Authors want a desktop app, including on macOS
You want vendor support for a self-hosted server, or you are invested in Salesforce

Choose Apache Superset if

Your team writes SQL and wants dashboards on top of saved queries
Your data is already in a warehouse that can serve interactive queries
You have many viewers and want to avoid per-viewer licences
You want to embed dashboards in a product without per-viewer fees
You can run Kubernetes services yourself, or will use Preset

When neither is right

You want open source BI that is easier for business users and simpler to host: see Tableau vs Metabase and Metabase vs Apache Superset.
You are a Microsoft organisation that wants a managed service with a semantic model: see Power BI vs Tableau and Power BI vs Apache Superset.
You need governed metrics in version-controlled code: see Tableau vs Looker.
You need scheduled tabular reports rather than dashboards: see the best SQL reporting tools.

Final recommendation

Bottom line

Tableau is the stronger exploration tool: canvas-based analysis, LOD calculations, extracts and a desktop app on two platforms, paid for per role. Apache Superset is the stronger fit when SQL is your team's working language and the warehouse can carry the load, because it costs nothing to license and keeps logic in SQL, but you take on its operation or pay Preset. Before choosing Superset, check that your database can serve dashboard queries live; before choosing Tableau, get a quote that covers every viewer. For other open source options, see the best open source BI tools.

Frequently asked questions

Is Apache Superset a free alternative to Tableau?

Superset is free under the Apache License 2.0, so it is often considered one. It covers dashboards and SQL-based exploration well, but it has no extract engine, no desktop app and a smaller set of calculation features than Tableau, and you must host it or pay Preset.

Can I write SQL in Tableau?

Yes, as custom SQL on most database connections. Tableau documents that custom SQL must be a single SELECT statement and is wrapped in an outer query, and that parameters can replace only literal values. Most analysis in Tableau is done on the canvas and in calculated fields rather than in SQL.

Does Superset have Tableau-style extracts?

No. Superset queries the connected database for every chart, with optional result caching through Redis. If the source is slow, teams usually move data into a warehouse or a reporting table first.

Can I self-host Tableau for free?

No. Tableau Server runs on your own Windows or Linux servers, but users still need Creator, Explorer or Viewer licences. Tableau Public is free but publishes your data openly. Superset is the free self-hosted option of the two.

How much does Tableau cost?

Tableau lists per-role prices on its pricing page, which we could not load to confirm in October 2026, so we do not quote them. Total cost depends on your number of Creators, Explorers and Viewers, the edition and any negotiated terms.

Sources

Checked October 2026.

How we research comparisons: our editorial method.