Quickstart

Build, test, and deploy your first Connector SDK connector using an AI coding agent.

Prerequisites

A Fivetran account with a configured destination
A supported Python version
An operating system supported by the Connector SDK
Connector SDK installed
FIVETRAN_API_KEY set as an environment variable. The value must be your base64-encoded Fivetran API key.
An AI coding agent of your choice installed
The Fivetran Connector SDK plugin for that agent installed

Open your agent in a new project directory

In your terminal, create a new directory and start your agent. For example, with Claude Code:

mkdir my-connector && cd my-connector
claude

Build the connector

Run the build skill:

/fivetran-connector-sdk:build-connector

When the agent asks what you want to build, describe your connector:

Build a connector for the OpenFDA Drug API (https://open.fda.gov/apis/drug/).
Use incremental sync with a date cursor. Replicate the drugsfda endpoint.
Limit to 10 records per endpoint during testing.

The agent:

Checks whether Fivetran already has a managed connector for your source. If one exists, it tells you. You can ask it to proceed anyway if you need a custom connector.
Researches the FDA Tobacco API and asks clarifying questions about authentication, endpoints, and primary keys.
Scaffolds the project using fivetran init and generates connector.py, configuration.json, and requirements.txt.

Enter your credentials

When prompted, open a new terminal window and run the command the agent provides:

python <plugin-tools-path>/enter_configuration.py configuration.json

The script prompts you for each field in configuration.json and encrypts every value locally so the agent never sees your plaintext credentials. Return to the agent chat and confirm when done.

This encryption is for local development only. After deployment, Fivetran stores your configuration using its own credential security. See Secret management.

Review test results

The agent runs fivetran debug and reports the results:

Tables synced: drug_drugsfda (10 rows)
Checkpoint: created

If errors occur, the agent fixes logic errors automatically or tells you what to check for infrastructure or credential issues.

Evaluate before deploying

Before deploying, run the evaluate-connector skill to check for SDK compliance, security, and reliability issues:

/fivetran-connector-sdk:evaluate-connector

The agent analyzes connector.py and produces a scored report. Fix all REQUIRED issues before deploying. See Evaluate connector quality.

Deploy

Once the connector passes evaluation, run the deploy-connector skill:

/fivetran-connector-sdk:deploy-connector

The agent deploys to Fivetran and gives you a link to your new connection. The connection starts paused. Confirm when you are ready to start syncing data.

Starting the initial sync consumes Monthly Active Rows (MAR).

Start and verify the initial sync

Open the connection using the URL returned by the agent, or find it on the Connections page in your Fivetran dashboard. Start the initial sync by clicking Start Initial Sync.

After the sync completes, verify that a schema and tables are created in the destination and records have loaded from the FDA Tobacco API.