AI that coordinates your monitoring

Know why it went down — not just that it did

Local statistics work with no API key. Add Claude, OpenAI or any compatible endpoint for incident summaries, root-cause hints, forecasts and postmortems.

MonitorForge sends the AI only check statistics — never page content. Monitor data is treated as untrusted input, and every call counts against a daily quota you control.

Collect the facts

Every check stores status, timing breakdown and errors. Local statistics find trends and outliers with no API key.

Explain the incident

On recovery the AI writes a short summary: what failed, how long, and the likely cause.

Act early

Pro forecasts latency drift and runs a 15-minute burst of faster re-checks when something looks wrong.

What the AI can do

  • Overview: p95 latency, trend and outliers across all monitors.
  • Incident summary: written automatically when a monitor recovers.
  • Root cause (Lite): hints from timing breakdown and error text.
  • Optimize, forecast, postmortem (Pro): capacity forecasts and shareable postmortems.
  • Adaptive re-checks (Pro): a 15-minute burst at your fastest interval when something looks wrong.

Add AI to your monitoring

Bring your own key. Nothing runs unless you turn it on.