Supported warehouses
Replit supports connections to:- BigQuery
- Databricks
- Snowflake
Admin setup
Administrators must configure warehouse connectors before team members can use them. The setup process involves configuring an OAuth application in your warehouse provider and adding those credentials to Replit.Prerequisites
- Enterprise plan
- Admin access to your Replit organization
- Ability to create OAuth applications in your warehouse provider (or access to credentials from your IT/Data team)
Configuration steps
1
Navigate to Integrations
Go to your organization’s settings and select the Integrations tab.
2
Enable Connector
Select the warehouse you want to connect (BigQuery, Databricks, or Snowflake).
3
Enter Credentials
Provide the Client ID and Client Secret for the OAuth application you created in your warehouse provider.
4
Configure Access
Use Role-Based Access Control (RBAC) to specify which members or groups can use this connector.
Warehouse-specific configuration
Databricks
Databricks
Admins can configure Databricks connections using one of two methods: User-OAuth or Service Account.Option 1: User-OAuthThis method involves minting OAuth application tokens within Databricks.Option 2: Service AccountThis method involves creating a service account (Service Principal) in Databricks to connect to Replit.The service account is shared among everyone given permission to use that integration. The permissions granted to the service account in Databricks will flow to all users enabled for the integration. Scope service accounts to READ-only access on the data when possible.Managing Access GranularityTo differentiate access levels (for example, restricting specific tables to different teams):
- Create multiple Service Principals in Databricks
- Assign specific permissions to each Service Principal
- Create separate integrations in Replit for each Service Principal
- Open access to each integration only for the allowed individuals or groups
Builder access and login
Once an admin has enabled a connector and granted you access, you can connect to the warehouse.Connecting to a warehouse
When you ask Agent to use a warehouse (for example, “Query the Snowflake database…”), or when you manually add the integration, you receive a login prompt. Each warehouse requires specific information during the login or connection process:BigQuery
BigQuery
Required at login:
- Project ID: You must specify the Google Cloud Project ID you want to access.
Databricks
Databricks
Required information:
- SQL Warehouse: The specific SQL Warehouse compute resource to use.
- Account URL: Your Databricks account URL.
Snowflake
Snowflake
Required information:
- Account ID: Your Snowflake Account ID.
Building with warehouse data
After connecting, you can ask Agent to build applications that use your data. Agent can:- Build internal tools that fetch and display live data
- Create dashboards with charts and visualizations backed by your warehouse
- Generate SQL queries to power your application’s backend
- Explain schema and table structures to help you understand what to build
While Agent can answer ad-hoc questions about your data, the primary purpose of Warehouse Connectors is to enable Agent to build functional applications that leverage your organization’s data. Warehouse queries are executed directly against your instance, so make sure your OAuth scopes and database user permissions allow the necessary read and write operations.
Related documentation
- Connectors overview — Learn about all integration types
- Connectors for Organizations — Centralized connector management for Teams and Enterprise