Help with Lead Qualification Workflow

Questions
5

I agree with Anshul’s Analysis. This is not a Telnyx or Vapi authorization issue. It’s an n8n webhook response-mode mismatch.

The webhook does reach n8n (executions start), but some executions never return an HTTP response. When that happens, n8n marks the run Failed, the webhook handshake isn’t closed, and Telnyx/Vapi treat the event as failed or retry it. Only executions that reach a response succeed — which is why it “sometimes works.”

Telephony platforms expect a fast, deterministic 200 OK and don’t wait for downstream workflow logic.

Fix: Pick one and make it deterministic

Recommended for telephony — Respond Immediately

  • Webhook node: Respond = Immediately

  • HTTP code: 200

  • No Respond to Webhook node

  • Downstream logic runs async

Alternative — Using Respond to Webhook

  • Webhook node: Using Respond to Webhook

  • Add a Respond to Webhook node

  • Every execution path, including error paths, must reach it

  • Any missed path causes intermittent failure

Do not mix the two patterns.

Sanity check

Outside of N8N call the Production webhook URL used by Vapi:

curl -i -X POST https://<n8n-cloud-url>/webhook/<path> \
  -H "content-type: application/json" \
  -d '{"test":"ping"}'

You should get an immediate 200 OK and no failed execution.

References