Connecting Databricks to your app

Connect Databricks to your app to run notebooks, query data, and manage pipelines.

Connect your app to Databricks to run notebooks, execute SQL, and manage data pipelines, then build dashboards, reports, and AI apps on top of your lakehouse data.

You need a Builder plan or higher to use connectors in your app.

---

What you can do

Turn your data and jobs into actions and insights in your app.

Run notebooks and jobs

Trigger Databricks notebooks and data pipeline jobs from your app.

Execute SQL

Run SQL against your lakehouse and SQL warehouses, and surface the results in your app.

Build analytics dashboards

Turn Databricks queries into charts, tables, and dashboards your team can read.

Example prompts

Describe the notebook, job, or query, what triggers it, and any dynamic fields in the AI chat. For example:

Run a Databricks notebook to process today's data.

Execute a SQL query on my Databricks lakehouse.

Trigger a data pipeline job in Databricks.

Fetch results from a Databricks SQL warehouse.

Show the status of my latest Databricks jobs in a dashboard.

---

Setting up OAuth in Databricks

You need a Databricks account admin to create a custom OAuth application before you connect Databricks to Base44.

1. Get the Base44 callback URL

   Get the Base44 callback URL from Registering callback URLs.

2. Create the Databricks OAuth application

   Follow Databricks’ guide to creating a custom OAuth application.

   In Databricks, add that callback URL as a redirect URI and select All APIs.

3. Generate your credentials

   Click Generate a client secret, then copy the Client ID and Client Secret. Databricks shows the Client Secret only once.

---

Connecting Databricks

Connect Databricks so your app can query your data.

Before you begin:

• Databricks OAuth app: Have a Databricks account admin create a custom OAuth application and generate a client secret. Keep its Client ID and Client Secret. See Setting up OAuth in Databricks.
• Databricks Workspace URL: Have the URL of your own Databricks workspace ready. Copy it from your browser while you’re signed in to that workspace.
• Base44 workspace access: You must be a workspace owner or admin to register the connector in Base44.
• Databricks access: Each person who connects needs access to the Databricks workspace resources your app uses.

To add Databricks from the Connectors page:

1. Click Connectors in your workspace sidebar.
2. Click Add connector.
3. Search for Databricks and click it.
4. Enter a Connector name.
5. Enter the Client ID and Client Secret from your Databricks OAuth app.
6. Keep the defaults under Scopes. Select sql if your app runs SQL. See Scopes and permissions.
7. Enter your own Databricks Workspace URL under Connection Configuration, for example https://dbc-xxxxxxxx-xxxx.cloud.databricks.com. It ends in .databricks.com, .azuredatabricks.net, .cloud.databricks.us, or .cloud.databricks.mil. Copy it from your browser, without a path or query string, and don’t enter databricks.com or accounts.cloud.databricks.com.
8. Click Add Connector.
9. Open your app in the app editor.
10. Ask the AI chat to add a Databricks feature.
11. Authorize the account when prompted.

Your app can use one shared Databricks connection or let each person connect their own account. Tell the AI chat which you want.

---

Scopes and permissions

Your Databricks permissions determine which data and workspace resources your app can use.

1. Base44 adds

   • all-apis: Calls Databricks APIs
   • offline_access: Refreshes the connection without another sign-in
2. You configure

   • sql: Runs SQL queries on Databricks SQL warehouses. Select it under Scopes if your app runs SQL.
   • Leave Additional scopes empty unless your app needs more.

Access: Databricks limits the connection to the connected identity’s existing workspace permissions.

---

Using Databricks in a backend function

Once Databricks is connected, your backend functions can retrieve an access token and call Databricks APIs. Base44 handles the OAuth flow and credential storage.

Shared connector

Use getWorkspaceConnection(connectorId) when your app uses one shared Databricks identity.

import { createClientFromRequest } from "npm:@base44/sdk@0.8.44";

Deno.serve(async (req) => {
  const base44 = createClientFromRequest(req);
  const connectorId = "YOUR_CONNECTOR_ID";

  const { accessToken, connectionConfig } =
    await base44.asServiceRole.connectors.getWorkspaceConnection(connectorId);

  const baseUrl = connectionConfig.host.startsWith("https://")
    ? connectionConfig.host
    : `https://${connectionConfig.host}`;

  // Use accessToken and baseUrl to call the Databricks REST API.
  return Response.json({ workspaceUrl: baseUrl });
});

App user connector

Use getCurrentAppUserConnection(connectorId) when each person connects with their own Databricks identity.

import { createClientFromRequest } from "npm:@base44/sdk@0.8.44";

Deno.serve(async (req) => {
  const base44 = createClientFromRequest(req);
  const user = await base44.auth.me();

  if (!user) {
    return Response.json({ error: "Unauthorized" }, { status: 401 });
  }

  const connectorId = "YOUR_CONNECTOR_ID";
  const { accessToken, connectionConfig } =
    await base44.asServiceRole.connectors.getCurrentAppUserConnection(connectorId);

  const baseUrl = connectionConfig.host.startsWith("https://")
    ? connectionConfig.host
    : `https://${connectionConfig.host}`;

  // Use accessToken and baseUrl for the signed-in person's Databricks requests.
  return Response.json({ workspaceUrl: baseUrl });
});

Connectors (developer docs)

Retrieve a connection, read its details, and call the service API from a backend function.

---

Troubleshooting

Check your workspace URL and Databricks permissions if the connector can’t connect or run queries.

The workspace URL is rejected

Your workspace host has to end in .databricks.com, .azuredatabricks.net, .cloud.databricks.us, or .cloud.databricks.mil. Enter it on its own, with no path, for example adb-123456789.10.azuredatabricks.net.

Queries fail with a permissions error

Each person who connects needs CAN_USE on the SQL warehouse your app queries, or Databricks returns PERMISSION_DENIED. For an app user connection, Unity Catalog applies per person, so the same query can return different rows for different people, or TABLE_OR_VIEW_NOT_FOUND for someone without access to a table.

A query fails without a warehouse ID

Databricks has no default SQL warehouse, so every query needs a warehouse ID. Find it in Databricks under SQL → SQL Warehouses, then open the warehouse and copy its ID.