Condit
condit/connectors/databricks

Chat with your Databricks data in plain English

Ask your Databricks warehouse questions in plain English — Condit writes the SQL, runs it, and explains the result, without routing through Condit’s own servers.

↓ Download Condit

Why Condit for Databricks

  • Built around how Databricks actually behaves. Large workspaces can have enough catalogs and schemas that eagerly scanning all of them causes multi-minute hangs in some tools. Condit lists every catalog immediately, fully loads the one you’re working in, and loads others’ schemas and tables on demand as you need them.
  • Fast catalog browsing. Uses Databricks’ native SHOW CATALOGS / SHOW SCHEMAS / SHOW TABLES instead of information_schema scans.
  • Connection pooling that stays alive. Condit recycles and pings its Databricks connection in the background, so you’re not paying reconnect overhead on every query in a session.
  • Unity Catalog-aware. Row counts, primary keys, and foreign keys are read straight from Unity Catalog metadata.
  • Token-based auth. Personal Access Token or OAuth.

How it works

  1. Download Condit and add a Databricks connection — workspace host, SQL warehouse HTTP path, and a Personal Access Token or OAuth login.
  2. Condit shows you every catalog in your workspace right away, and loads the one you’re using in full.
  3. Ask your question in plain English.
  4. Condit writes SQL scoped to your catalog/schema and runs it on your warehouse.
  5. Review the results, or refine the query yourself in the SQL editor.

Example

You → how many orders shipped last week, by region?
SQL → SELECT region, COUNT(*) FROM sales.gold.orders WHERE ship_date >= … GROUP BY region

Generated using your actual catalog and schema names, not a placeholder.

FAQ

Does Condit support Databricks?

Yes — SQL warehouses via the official Databricks SQL connector, with PAT or OAuth auth.

Why does catalog browsing feel fast?

Condit avoids the eager, all-catalogs scan that causes slow hangs on large workspaces — it shows every catalog immediately and loads detail for others as you actually need it.

Does Condit run my SQL locally?

Queries run directly from your machine to your Databricks warehouse — never through Condit’s servers.

Does my data reach the cloud?

Only if you pick a hosted LLM — same model as every other connector; a local Ollama model keeps everything on your machine.

Can I use my own LLM?

Yes — OpenAI, Anthropic, Gemini, Groq, or fully local via Ollama.

Also connecting to Snowflake? See the Snowflake connector →

↓ Download Condit