Skip to content

Data Sources

Catalyst can connect to your databases to enable AI-powered data analysis. Query your data directly and use results in prompts and workflows.

  • PostgreSQL
  • MySQL
  • Oracle
  • SQLite
  • SQL Server
  1. Navigate to Data Sources from the sidebar menu
  2. Click Add Data Source
  3. Fill in the connection details:
    • Name - A friendly name for this connection
    • Type - Select your database type
    • Host - Database server hostname or IP
    • Port - Database port (defaults vary by type)
    • Database - Database name
    • Username - Database user
    • Password - Database password
  4. Click Test Connection to verify
  5. Click Save
  • Credentials are encrypted before storage
  • Connections use read-only mode where possible
  • All queries are validated to prevent data modification
  • Only SELECT statements are allowed

Once connected, you can use your data sources in:

Select a data source in a prompt template to query data and include results in your AI prompts. See Prompts for details.

Add database query nodes to your visual workflows to fetch and process data. See Workflows for details.

For databases running on your local machine or local network, ensure:

  • The database is accessible from your browser
  • Firewall rules allow the connection
  • The database accepts connections from external hosts

Catalyst supports Tailscale for secure connections to private databases:

  1. Navigate to Network from the sidebar
  2. Accept the device invite to join your Tailscale network
  3. Your data sources can now use Tailscale hostnames

This allows you to securely connect to databases that aren’t publicly accessible.

In addition to databases, Catalyst supports file-based data:

  • CSV files - Comma-separated values
  • JSON files - JSON data files
  • Markdown files - Converted via MarkItDown

Upload files directly when configuring a data source.

  • Verify the hostname, port, and credentials
  • Check that the database server is running
  • Ensure firewall rules allow the connection
  • For cloud databases, check IP allowlists
  • Only SELECT queries are supported
  • Ensure you have read permissions on the tables
  • Check SQL syntax matches your database type