Introducing the AI Copilot

Status pages generate a lot of data. Nobody reads it.

You've got 5 monitors, 3 components, a handful of resolved incidents from last week, and 99.7% uptime. Good numbers. But when someone asks "how are things looking?" you end up digging through the dashboard to piece together an answer.

We figured the AI should do that for you.

What's new

Three things shipped today.

Dashboard summaries

Your dashboard now has an AI Summary card at the top. It looks at your monitors, incidents, and uptime data, then gives you a one-paragraph status overview. Something like "All 5 monitors are healthy. No incidents in the past 7 days. Overall uptime: 99.9%."

On the default "Suggest & Confirm" mode, you click Generate to trigger it. If you switch to Semi-Autonomous or Full Autopilot, it generates automatically when you load the page.

Summaries are cached for 5 minutes so you're not burning tokens on every page refresh.

Public status page summary

There's a new toggle in AI Settings called "Public AI Summary." Turn it on and your public status page gets a one-line AI-generated sentence at the top. "All systems operational" or "Some services are experiencing issues" depending on what's actually happening.

It's capped at 100 characters. No fluff, no jargon.

Weekly reliability digest

Every Monday at 9 AM UTC, Chirp can email you a summary of your week. Uptime percentage, incident count, average resolution time, most affected component, and a comparison to the previous week.

You enable it with a toggle in settings. The email includes both the raw stats and an AI-written paragraph putting the numbers in context.

Three levels of AI autonomy

Instead of a single on/off switch, you pick how involved the AI should be.

Suggest & Confirm is the default. AI doesn't do anything until you ask. Dashboard summaries need a click. Incident suggestions wait for your approval.

Semi-Autonomous lets AI auto-generate summaries and descriptions. Incidents and notifications still need you to say yes.

Full Autopilot (Pro only) hands over the whole incident lifecycle. AI detects issues, creates incidents, writes updates, and resolves them. You step in for edge cases.

You can change this anytime in your page settings under "AI Settings."

Usage tracking

Every AI generation gets logged with token counts and estimated cost. You can see your 30-day usage right in the AI Settings section: total generations and estimated cost in dollars.

We use DeepSeek as the primary provider with Anthropic as fallback. If both are unavailable, a template generates the text with zero token cost. You're never left without a summary.

Getting started

All of this is live now. Summaries, public status sentences, and the weekly digest work on every plan, including Free. Full Autopilot mode is the only Pro-exclusive feature.

Head to your dashboard, open any status page, and you'll see the AI Summary card. Settings are under the "AI Settings" section at the bottom of your page settings.

If you have feedback or ideas for what the AI should do next, find me on X.