Databricks MCP server
Run Databricks SQL, Genie spaces, and Unity Catalog functions from your AI agents.
About
Connect Tracecat to Databricks to read and reason over the data already governed by your lakehouse for SOC analysts and detection engineers running security pipelines on Delta tables. You can query Databricks SQL for historical events from a workflow with no separate credentials to manage. From there, ask a Genie space a natural-language question, run Vector Search over indexed data, and invoke Unity Catalog functions with every call attributed to the connected principal in your Databricks audit logs.
Setup
- 1
Sign in with OAuth
You'll authorize Tracecat to access Databricks on your behalf. No API keys to manage.
- 2
Select the
Databrickstile in the Tracecat MCP catalogOpen the
MCP catalogin your workspace, select theDatabrickstile, and complete the OAuth flow. - 3
Enable
Databricksin your agentIn your
ai.agentaction orAgents→toolstab, selectDatabricksfrom theMCP integrationsdropdown.
Tools
Vector Search | Query Databricks Vector Search indexes for semantic retrieval over indexed data. |
Genie Spaces | Ask natural-language questions against a Genie space and get SQL-grounded answers. |
Databricks SQL | Execute SQL against a Databricks SQL warehouse through the AI Gateway. |
Unity Catalog functions | Invoke Unity Catalog functions registered as MCP tools, governed by Unity Catalog grants. |