Testing a webhook workflow end to end
From an inactive draft to a verified live run: trigger for real, watch the execution, and bring failures back to the chat.
Deploy inactive first
Finish node configuration — for example a Webhook path of register-user and a Postgres insert query — save every node, pass preflight, and deploy an inactive draft. Inactive means testable without consequences.
Trigger it for real
Open the workflow in n8n and send a real request to the webhook URL shown on the trigger node:
curl -X POST https://your-n8n-host/webhook/register-user \
-H "Content-Type: application/json" \
-d '{"email": "test@example.com"}'Replace the host with your own instance address. Never test against localhost URLs — remote instances cannot reach them.
Read the execution, fix, repeat
Open Executions in n8n to see each run with per-node inputs and outputs. When a node errors, copy the failing input or message back into Copilot, fix the parameter, save, and deploy again. Two or three of these loops is a normal, healthy pace — not a sign something is wrong.
Next guide
Avoiding duplicate deployments
Questions? Contact Support.