Voice & text-to-speech calls

Some messages are too important to sit unread in an inbox. The voice channel calls the recipient and reads your message aloud with natural text-to-speech — ideal for critical alerts, on-call escalation, appointment reminders for less digital audiences, and OTP fallback.

Setup

Open Settings → Channels → Voice and connect your telephony credentials (Twilio-compatible). One-time setup via PUT /v1/channels/voice.

Sending a call

Voice uses the exact same send endpoint as SMS and email — only the channel differs:

curl -X POST https://api.wezend.com/v1/messages/send \
  -H "X-API-Key: $WEZEND_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "to": "+4512345678",
    "channel": "voice",
    "message": "This is an automated alert from Acme. Your server EU-1 is unreachable. Press any key to acknowledge."
  }'

What you get back

Call status flows through the same lifecycle as any message (queued → sent → delivered/failed), where delivered means the call was answered. Delivery webhooks fire exactly like for SMS, so your escalation logic ("no answer in 2 minutes → call the next person on rotation") is a small journey rather than custom code.

When to prefer voice

• OTP fallback — if the SMS wasn't delivered within N seconds, call instead (built as a two-step journey with a wait_event node).
• Critical ops alerts — voice interrupts; push and email do not.
• Audiences that don't read SMS — reminders for healthcare and public-sector use cases.