AgentPulse

Find out your agent stopped running from an email, not from a gap in the data three weeks later.

What you get

• One ping URL per loop: add a line of curl at the end of the run and you are monitored
• Silence is the alert — set how often the loop should check in and how late is still acceptable; one email when it goes quiet, one when it comes back
• Report failures explicitly: ping /fail with the exit code or error and get told immediately, even though the loop did run
• Live dashboard: status, last check-in, next due countdown, 48-run sparkline and a full event timeline per monitor
• Copy-paste integrations for cron, bash, Node, Python, PowerShell, n8n, GitHub Actions and Claude Code Stop hooks
• Alerts to any address you choose, per monitor; pause during maintenance; rotate a token if it leaks
• Up to 25 monitors, check-ins from every minute to weekly

Questions