AI Connector Agent Setup Guide Private Preview

Follow this guide to quickly build your connector with the AI Connector Agent and connect your data source to Fivetran.

Prerequisites

To use the AI Connector Agent, you need a Fivetran account and a user role with permissions to create a connection.

Setup instructions

Specify source details

1. Enter the name of your data source.

2. Enter the URL of the data source API documentation.

   You must enter the API documentation homepage URL, not a specific endpoint or subpage. For example, https://dev.frontapp.com/docs/welcome or https://app.delighted.com/docs/api. Correct URLs help the agent accurately discover and interpret the complete set of available endpoints.

3. (Optional) To help the agent identify and extract the API endpoints more accurately, set the Upload API documentation toggle to ON and upload one of the following files:

   • OpenAPI specification (YAML or JSON)
   • Postman collection (JSON)
   • API documentation (PDF)

4. Read the Fivetran terms and conditions and select the I consent to Fivetran's AI Connector Agent terms checkbox.

5. Click Generate. The agent starts scanning the API documentation and building the connector.

Depending on the source complexity and number of endpoints, the agent may take anywhere from a few minutes to over an hour to list all the detected endpoints.

Review API endpoints

1. In the Review endpoints page, review the list of entities and the associated API endpoints that the agent extracted from the API documentation. By default, all entities are selected.

   The Table column lists the tables that will store data from each of these API endpoints.

2. Clear the checkboxes for any entities you do not want to sync.

3. Click Continue.

Set up connection

1. In the connection setup form, fill in all the fields.

   The agent automatically generates a unique name for the Destination schema. You can retain the default name or enter a name of your choice. However, you cannot change the schema name after you click Save & Test.

2. Click Save & Test. Fivetran tests and validates the connection.

3. Once the connection tests pass, click Continue.

4. Choose one of the following options:

   • Start syncing all my data now: Starts the initial sync immediately and syncs all data from the API endpoints you selected. This option runs a historical sync and captures all the data without customization. To start the sync immediately, click Start Initial Sync.

   • I need to customize data before syncing: Lets you review and customize the connection schema before starting the initial sync. You can block specific tables or columns, hash sensitive data, and define how to handle future changes to the source schema. To customize the schema before syncing your data, do the following:

     1. Review the connection schema and clear the checkboxes for any tables or columns you want to exclude. By default, all tables and columns are selected.

     2. Click Save & Continue.

     3. Choose how to handle future changes to the source schema.

     4. Click Continue.

     5. Click Start Initial Sync.

   • I'll sync later: Defers the initial sync. Choose this option if you are not ready to start syncing data immediately. To start the sync later:

     1. In the Fivetran dashboard, go to the Connections page.

     2. Select the connection that the agent created.

     3. Click Start Initial Sync.

        To review and customize the connection schema before syncing your data, click Customize data before syncing before you start the initial sync.