Skip to main content

​
Getting Started

Julius connects directly to your Databricks SQL warehouse, allowing you to analyze your data using natural language; no SQL knowledge required.

Create New Data Connector

Navigate to the Data Connectors page and click “Create new Data Connector”, then select Databricks.

​
Connection Setup

​
Required Fields

All fields marked with an asterisk (*) are required. MFA_TYPE is optional.
​
Connection Name
string
required
A friendly name to identify this Databricks connection in Julius.Example: “Prod Databricks Warehouse” or “Marketing Analytics Workspace”
​
SERVER_HOSTNAME
string
required
The Databricks workspace hostname.Example: adb-1234567890123456.7.azuredatabricks.net
​
HTTP_PATH
string
required
The HTTP path for your Databricks SQL warehouse or cluster endpoint.Where to find it: In Databricks, go to your SQL warehouse, click Connection details, and copy the HTTP Path value.
​
ACCESS_TOKEN
string
required
A Databricks personal access token (PAT) or equivalent credential with permission to query the warehouse.Security: Julius encrypts and securely stores this token.
​
CATALOG
string
required
The Databricks catalog Julius should use by default.Example: main or your organization’s primary catalog.
​
SCHEMA
string
required
The default schema within the catalog that Julius will query.Example: analytics, public, or another schema that contains your tables.
​
MFA_TYPE
string
Optional multi-factor authentication type if your organization requires MFA for Databricks access.Most users: Can leave this blank unless specifically required by your security policy.

​
Network Configuration

Test Your Connection

  1. Click “Add Connection” to test your configuration
  2. Julius validates your credentials and network access
  3. If successful, your connector is ready to use!
  • ✓ Host and port are correct
  • ✓ Database allows external connections
  • ✓ Username and password are accurate
  • ✓ User has permissions for the specified database

​
Start Analyzing Your Data

Once connected, simply ask Julius about your data in plain English:

Query Data

“Show me all customers from the users table”

Calculate Metrics

“What’s the average order value this month?”

Create Visualizations

“Create a chart of sales by region”

Analyze Trends

“How has revenue changed over the past 6 months?”
No SQL required! Julius automatically writes and executes queries based on your natural language requests.

​
Security & Compliance

SOC 2 Type 2

Certified to meet strict security standards

End-to-End Encryption

All credentials and data encrypted in transit and at rest

Zero Data Storage

Julius only queries what you request—never stores your data

Security & Trust Center

Learn more about how Julius protects your data and maintains compliance

​
Best Practices

Create a dedicated Databricks service principal or user with read-only access to the catalogs and schemas Julius needs. Avoid using admin tokens or workspace-wide admin accounts.
Verify your connection works correctly using a development or staging database before connecting to production.
Work with your database administrator to ensure proper security policies and performance guidelines are followed.
Keep an eye on database load when running complex queries through Julius, especially on production systems.

​
Troubleshooting

Possible causes:
  • Incorrect host or port
  • Database doesn’t allow external connections
  • Database is offline or unreachable
Solution: Verify all connection details and check firewall settings
Possible causes:
  • Wrong username or password
  • User lacks database permissions
  • Database name misspelled
Solution: Double-check credentials and verify user permissions
Possible causes:
  • User lacks read permissions on tables
  • Connected to wrong database
  • Tables don’t exist in specified schema
Solution: Verify user has SELECT permissions and correct database is specified

​
Need Help?

Contact Support

Questions? Reach out to team@julius.ai for assistance