WhatsApp Outbound Call Webhooks for Standard Workspaces
A standard UnifyPort workspace can prepare one-to-one outbound WhatsApp calls. If your Webhook endpoint only subscribes to ["*"], messages can keep arriving while outbound call callbacks are absent. Add call.answered and call.ended by name to receive both events.
The create-call reference covers voice and video calls started by a person through UnifyPort’s unofficial interface. call.incoming remains reserved for future inbound calling.
1. Check the subscription rule
| Workspace | Subscription | Outbound call events |
|---|---|---|
| Standard | ["*"], [] or an omitted list | Regular events only |
| Standard | Explicit names in subscribed_events | Each listed outbound call event |
| Beta | Explicit names or wildcard, empty or omitted list | Matching outbound call events |
An explicit event name receives the corresponding event when the endpoint is eligible. In a standard workspace, including * alongside the two names preserves wildcard coverage for regular events. Explicitly subscribing to reserved call.incoming in a standard workspace returns 403 beta_feature_not_enabled (41002).
2. Read the endpoint before changing it
Use Get webhook endpoint to confirm the destination URL, status, current subscription and retry policy. Check whether the endpoint belongs to the workspace or a particular messaging account. The update keeps the existing account binding.
Update webhook endpoint replaces the configuration. Prepare the full desired settings before sending it. Preserve the current URL, status and retry policy. If signing is enabled, retrieve the existing secret from your own secure configuration. The read response exposes signing_enabled, and the secret stays with your application. Omitting or emptying signing_secret disables signing. Omitting the retry policy restores the default of three retries after the initial attempt.
3. Add the events while preserving message delivery
For an endpoint that currently uses wildcard coverage, send the following body to PATCH /v1/webhook-endpoints/{endpoint_id}. The example assumes an active, signed endpoint that already uses three retries. Substitute your current URL, status, secret and retry settings. Keep API keys and signing secrets on the server.
{
"url": "https://support.example.com/webhook",
"status": "active",
"subscribed_events": ["*", "call.answered", "call.ended"],
"signing_secret": "<EXISTING_WEBHOOK_SIGNING_SECRET>",
"retry_policy": { "max_attempts": 3 }
}
For an endpoint with an explicit list, copy the current event names and add the two call events. For example, if the existing list contains only message.received and message.read, the resulting value is shown below. Use this array as the subscribed_events field in the full update body.
["message.received", "message.read", "call.answered", "call.ended"]
After updating, read the endpoint again. Compare URL, status, subscription, signing state and retry settings with your intended configuration. Confirm that the receiver can route both call events and existing message events.
4. Verify with one new outbound call
- Use a test conversation selected from the same authorized WhatsApp messaging account. Prepare a call with
POST /v1/callsand save the returnedcall_idwith your application task. - A
201response withstate: preparingconfirms the link is ready. The operator opens the privatejoin_urland clicks Start call beforeexpires_at. - Once the test call connects and ends, look for the answered event and the ended event. Match the event’s
account_idanddata.call.idto your saved record. - Send a test message through the existing workflow and confirm its expected event still reaches the receiver.
For signed endpoints, verify the raw request body as described in the delivery and signature guide. Deduplicate by event id and handle arrival out of order. A late answered event can add evidence of connection while the call record stays ended. Message call records under message.* have their own purpose.
5. Diagnose gaps and retain a rollback path
If a callback is missing, first check the exact event names and the endpoint’s scope and active status. Then check receiver logs, HTTP responses and application storage. A 2xx acknowledges delivery. Verify persistence and task association separately. Delivery, retries and dead-letter replay use the current eligibility and subscription settings.
Keep the previous endpoint settings and its signing secret reference available for rollback. To roll back, submit the previous full configuration and verify message delivery again. If support is needed, share the endpoint ID, account ID, call ID, UTC times and redacted receiver results through technical support. Keep keys, secrets and full call links private.
Turn messaging integration into a stable product pipeline.
Start by sending through one API, then bring every inbound message back into your business system with standard events.