Data Sources

Data sources let you connect external platforms to Confident AI so you can automatically generate evaluation datasets from your existing documents and messages, rather than creating test cases manually.

Supported Data Sources

Create a Data Source

1. Navigate to Project Settings > Data Sources
2. Click New Data Source
3. Enter a unique Name for the data source
4. Select a Type from the dropdown (Google Drive, SharePoint, Slack, or Notion)
5. Click Save

Configure Credentials

After creating a data source, you need to provide credentials so Confident AI can access the external platform. Click into the data source to open its detail page and fill in the required fields.

Google Drive requires:

• Service Account JSON - Credentials from a Google Cloud service account
• Folder URL - URL of the Google Drive folder to sync

SharePoint requires:

• Tenant ID - Azure tenant identifier
• Client ID - Azure app registration client ID
• Client Secret - Azure app registration secret

Slack requires:

• Bot Token - Slack bot token (starts with xoxb-)

Notion requires:

• Integration Token - Notion integration secret token

Connect and Sync

Once credentials are saved, click Connect to establish the connection. After connecting:

• Manual sync - Click Sync to pull the latest data at any time
• Automatic sync - Configure a sync schedule to keep data up to date automatically

Sync Schedule

You can set the sync frequency to one of the following intervals:

• Every hour
• Every 2 hours
• Every 6 hours (default)
• Every 12 hours
• Every 24 hours

Toggle Sync Enabled on or off to control whether automatic syncing is active.

Use Data Sources for Dataset Generation

Once a data source is connected and synced, you can use it to auto-generate evaluation datasets. When creating a dataset, link a generation configuration to a data source and Confident AI will process the synced content to produce test cases (goldens) automatically.

Manage Data Sources

From the data sources list, use the action buttons on any row to Sync, Edit, or Delete a data source.

To disconnect a data source without deleting it, open the detail page and click Disconnect. This removes the connection but preserves the data source configuration.