Agent setup

Give your agents a way to reach a human.

When an agent finishes work, detects a risk, or needs approval, it can create a YBY Bell escalation chain instead of hoping someone checks a log.

Job finished

Notify a reviewer when a long-running build, report, scrape, or eval has completed.

Risk detected

Escalate certificate expiry, failed health checks, suspicious changes, or billing issues.

Approval needed

Ask a human to approve deployment, payment, message sending, or a customer-facing action.

API preview

A tiny call creates a full escalation chain.

The public API is not live yet, but this is the customer-facing integration model.

POST /v1/reminders
Authorization: Bearer YOUR_API_KEY

{
  "title": "Production deploy needs approval",
  "owner": "founder",
  "urgency": "critical",
  "channels": ["chat", "sms", "call"],
  "schedule": "now",
  "stop_when": "acknowledged"
}

Verified recipients

Agents can only reach approved users and teams, reducing the risk of accidental outreach.

Delivery timeline

Every nudge, escalation, call, and acknowledgement becomes part of the chain history.

Runs after the session

The reminder continues on YBY Bell even after the agent session ends or the laptop closes.