On-call voice escalator

When a Sentry alert fires, place a Twilio call to the on-call engineer and read the incident summary aloud.

Cost~$0.05 per call (under 1 min)

Operations

voice_call.createtext_generation.generate

Prerequisites

voice_call.create enabled. Daily request limit of 20 so an alert storm doesn't bankrupt you.
Webhook receiver for Sentry that knows your on-call rotation.

Walkthrough

1. Summarize the alert

Take the Sentry payload, run it through GPT-4o-mini to get a 25-word spoken summary.

2. Place the call

Twilio dials. The platform reconciles real call duration via webhook (DEC-027) and charges actual cost, not the upfront estimate.

bash# Place the call. Real duration is reconciled via webhook after pickup.
curl -X POST https://www.upivia.com/v1/service-requests \
  -H "Authorization: Bearer $AGENT_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "service": "voice_call",
    "operation": "create",
    "payload": {
      "to": "+15555550123",
      "from": "+15555550000",
      "say": "Sentry alert. API error rate 8 percent. Last deploy 14 minutes ago."
    }
  }'

3. Audit the spend

After the call completes Twilio posts back; the ServiceRequest moves to `executed` and the wallet is debited the real cost. View at /usage.

Next steps

Audit every call at /audit-logs, watch spend at /usage, and tune budgets per service on the agent's page.

Create an account →