Skip to main content

​
Databricks

In one line: Connect your Databricks workspace so your UnleashX agent can query your lakehouse, browse catalogs, schemas, and tables, and trigger jobs — all in plain language.

​
1. Overview

Databricks is a unified data and AI platform built on the lakehouse architecture. Your workspace holds the catalogs, schemas, and tables that make up your data estate, the SQL warehouses that run queries, and the jobs and pipelines that keep everything fresh. Once connected to UnleashX, your agent can run SQL statements against a SQL warehouse, list and describe catalogs, schemas, and tables in Unity Catalog, preview data, and check or trigger jobs. It works against your live workspace using a personal access token, so it acts with your own identity and permissions. Connecting Databricks to UnleashX turns your lakehouse into something you can ask questions of directly. Instead of opening a notebook to run a query or check a table, you ask your agent in plain language and it runs the SQL, reads the schema, or reports on a job for you.

​
2. What you’ll need

  • A Databricks workspace and your workspace URL (e.g. https://dbc-xxxxxxxx-xxxx.cloud.databricks.com or your Azure/GCP host).
  • A SQL warehouse (or cluster) you can run queries on, and its warehouse ID.
  • Permission (via Unity Catalog grants) to read the catalogs, schemas, and tables you want the agent to access.
  • A personal access token for your Databricks user (see Section 3).
  • An UnleashX account with access to Data Connectors.
You provide a workspace URL, a warehouse ID, and a personal access token. UnleashX uses the token as a Bearer credential on the Databricks REST API.
Databricks enforces Unity Catalog permissions. The agent can only query and read objects your token’s user has been granted access to. Ask a workspace admin to grant the needed USE/SELECT privileges if a query is denied.

​
3. Get your credentials

Databricks MCP authenticates with a personal access token (PAT) scoped to your user in a specific workspace.
1

Open your Databricks workspace

Sign in to your workspace and note its URL — you’ll provide this to your agent as the host.
2

Generate a personal access token

Click your username (top right) → Settings → Developer → Access tokens → Manage → Generate new token. Add a comment (e.g. “UnleashX”) and set a lifetime, then Generate. Copy the token now — Databricks won’t show it again.
3

Find your SQL warehouse ID

Go to SQL Warehouses, open the warehouse you’ll use, and copy its ID from the details panel (or the URL). The agent runs queries against this warehouse.
4

Paste the details into UnleashX

When connecting (Section 4), paste your workspace URL, warehouse ID, and personal access token. UnleashX sends the token as an Authorization: Bearer <token> header on every request.
A personal access token is as powerful as your account within its granted scopes. Store it securely, set an expiry, and revoke it if it’s ever exposed. Prefer a token tied to a service principal with least-privilege grants where possible.

​
Permissions this token grants UnleashX

​
4. Connect on UnleashX

1

Open your agent

Go to https://www.tryunleashx.com and open the agent you want to connect.
2

Open Data Connectors

Inside the agent, go to Data Connectors.
3

Find Databricks

Locate Databricks and click Connect (or Add / Configure).
4

Enter your workspace details

Paste your workspace URL, SQL warehouse ID, and personal access token.
5

Confirm success

Click Connect. UnleashX validates the token and the Databricks connector shows a Connected badge.

​
Use Databricks in a Workflow

Once connected, you can add Databricks to any automation from the Workflows builder. Its triggers and tools appear in the Apps panel, marked with an MCP badge.
1

Add a trigger or action node

Open Workflows → New Workflow. On the canvas, click + Add Trigger (or the + below any node) to add a step.Add a trigger on the workflow canvas
2

Pick Databricks from the Apps panel

In the Paths panel, open the Apps tab and select Databricks. Use the search box if you have many connectors.Choose your app from the Apps list
3

Choose a trigger or tool

Pick the action you want (for example, “run SQL query”). Configure its fields — required fields are marked with a red asterisk (*).
4

Add or select your account

Under Selected account, choose an already-connected account, or click Add Account to connect one now.
5

Save and test

Fill in the remaining fields and click Save. Use Test to verify the step, then toggle Publish when the workflow is ready.
The steps are the same for every connector. For the full workflow builder guide, see Using MCP in Workflows.

​
5. Available tools

*Read-only SELECT statements don’t change data. If you allow INSERT/UPDATE/DELETE/MERGE or DDL, Run Query can modify or drop data — restrict the token’s grants to read-only if you want the agent limited to queries.

​
6. Example usage

“How many orders did we ship last week? Query the sales.gold.orders table.” → Runs Run Query with a SELECT COUNT(*) … WHERE ship_date >= … against your warehouse and returns the number. “List the tables in the marketing schema and describe the campaigns table.” → Runs List Tables on marketing, then Get Table on campaigns to return its columns and types.

​
7. Permissions & data access

UnleashX can:
  • Run SQL statements on the warehouse you configured.
  • List and describe catalogs, schemas, and tables, and preview rows.
  • List and start/stop SQL warehouses.
  • List, inspect, and trigger jobs — all subject to your token’s Unity Catalog and workspace permissions.
UnleashX cannot:
  • Access catalogs, schemas, or tables your token’s user hasn’t been granted.
  • Query a different workspace than the one you configured.
  • Manage workspace users, tokens, or admin settings.
To disconnect: In UnleashX, open Data Connectors → Databricks → Disconnect. Then revoke the token in Databricks under Settings → Developer → Access tokens (delete the “UnleashX” token). Disconnecting and revoking removes access immediately.

​
8. Troubleshooting

For general MCP issues, see /mcp/integrations.

​
9. Frequently asked questions

Is my data stored by UnleashX? No. UnleashX runs queries and reads metadata live through the Databricks REST API per request; it doesn’t copy your tables. Can the agent modify or drop data? Only if your token has write/DDL grants and you run a statement that changes data. For a read-only agent, grant the token SELECT-only access. Does this work on AWS, Azure, and GCP? Yes — provide the workspace URL for your cloud. Azure Databricks and Databricks on GCP use the same REST APIs. Should I use a service principal instead of my own token? For production, yes — a service principal with least-privilege grants is safer than a personal token and isn’t tied to an individual user.

​
10. References