An integration that makes failure visible.

The operational problem

A timeout does not necessarily mean an operation failed. If a shipping provider created the shipment before the response was lost, a blind retry can create another shipment.

Who uses it

Shipping operators validate and queue orders, inspect exceptions, and process batches. A read-only reviewer can inspect the history.

The complete workflow

1. 01Preview and import orders from the defined CSV format.
2. 02Correct missing address information before queueing.
3. 03Choose a success, transient-failure, or ambiguous-timeout scenario.
4. 04Process the queue, then retry or reconcile the saved job.
5. 05Send simulated tracking events and inspect the resulting shipment history.

Decisions that matter

• The order reference prevents duplicate imports.
• A stable job key identifies the intended shipment across attempts.
• The timeout scenario first checks the existing simulated provider record.
• Duplicate tracking events have no additional effect; older events cannot regress delivery.

What you can inspect

Open your own demonstration workspace, change roles, complete the workflow, refresh the page, and inspect the saved activity history. Export the current workspace to review the records.

Implemented technology

React and TypeScript, built with Vinext, running on Cloudflare Workers. D1 stores workspace records and server-side sessions; R2 stores maintenance images. Workflow transitions and permissions are checked by the server.

Current limits

Local integration simulator only. ShipStation is not connected. No real labels, carrier requests, or shipments are purchased. Queue processing is initiated manually; background execution and external webhooks require a separate connected deployment.

Demonstration roles are deliberately selectable. They illustrate server-enforced permissions within an isolated sample workspace, not a production employee or supplier identity system.

Public project brief