Skip to content
Start Free

Databricks analytics, without the export

Databricks holds the lakehouse tables, and usually the model outputs written back beside them.

Start FreeBook a 30-minute demo

Free plan. No credit card required.

What teams ask of Databricks

Typed the way you would say them. Each answer comes back with its chart, its numbers and the SQL that produced them, so it can be checked before anyone acts on it.

  • “How do the model's predictions compare with what actually happened?”
  • “Does this metric come out the same on the gold table as it does on the silver one?”
  • “Which accounts does the churn model rank highest, and what do they have in common?”
  • “How much did last month's numbers move after the backfill?”

What it is combined with

The questions that matter rarely live in one system. An agent can hold more than one source: ask a question that spans them and Supaboard queries each one, then combines the results, so the comparison happens when the question is asked rather than in a pipeline project that has to be scoped first.

  • Snowflake — plenty of companies run both, and the question usually spans them.

How the connection works

Databricks is one of the 12 databases and warehouses you can connect yourself. It connects directly and is queried when a question is asked, with nothing copied ahead of time, so setting up a first analysis takes about 5 minutes.

Databricks connects with a personal access token and the HTTP path of a SQL warehouse, then you pick a catalog and a schema. Supaboard only reads: it does not run jobs, trigger pipelines or write anything back.

  • Database and warehouse queries are read-only. Changes to connected apps require your approval.
  • Credentials are encrypted at rest, scoped per workspace, never logged and never returned to a client.
  • On the Enterprise plan the whole platform runs inside your own VPC. Supaboard receives no platform data; inference follows the model endpoint you configure. See Enterprise.
  • A free plan with no credit card, then from $19 per editor per month — see pricing.

More sources

Other connectors with a guide:

Also in Databases & warehouses: PostgreSQL, MySQL, Microsoft SQL Server, Oracle, MongoDB, Supabase, Snowflake, BigQuery, AWS Redshift, ClickHouse, MotherDuck. See how files, databases and warehouses connect.

Weighing this against another tool? See the sourced comparison of 14 BI platforms, or what an AI data analyst does with a connection like this one.

Data warehouses

Ask your first question from Databricks

Start FreeBook a 30-minute demo

Free plan. No credit card required.

Settings