Install guide and config at curatedmcp.com

PagerDuty MCP: Let AI Handle Your On-Call Chaos

Operations teams spend too much time context-switching between incident dashboards, chat, and email. PagerDuty MCP is the official integration that gives Claude, Cursor, and other AI agents direct access to your incident response workflows—so they can create incidents, manage alerts, update on-call schedules, and keep responders aligned without human hand-offs.

What It Does

PagerDuty MCP bridges the gap between your AI agent and your incident management system. Instead of manually toggling between tools, you can now ask Claude to create an incident, check who's on-call, escalate a critical alert, or annotate a timeline—and the agent executes those commands directly in PagerDuty.

The server unlocks core capabilities: create and update incidents with custom urgency and priority levels, list active on-call users and rotation schedules, manage escalation policies and team assignments, query historical alerts and response timelines, annotate incidents with context or responder notes, and trigger incident actions programmatically from plain-language AI commands.

This is especially powerful for automation. When your monitoring system fires a critical alert, an AI agent can automatically open a PagerDuty incident, notify the right on-call engineer, and add diagnostic context—all in seconds. Or during an active incident, your agent can pull recent alert history, suggest escalation paths, and track response metadata without anyone leaving their terminal.

How to Install

npx -y @pagerduty/mcp

Add the server to your Claude Desktop config:

{
  "mcpServers": {
    "pagerduty-mcp": {
      "command": "npx -y @pagerduty/mcp",
      "env": {
        "PAGERDUTY_API_TOKEN": "your-api-token-here"
      }
    }
  }
}

You'll need a valid PagerDuty API token with appropriate scopes. Get yours from your PagerDuty account settings.

Real-World Use Cases

• Incident triage at scale: Ask Claude to review your last 50 incidents, identify patterns (e.g., recurring database timeouts), and propose new escalation policies without manually digging through the UI.
• Automated alert response: When a monitoring tool detects anomalies, an AI agent creates a PagerDuty incident, pulls the on-call roster, checks recent similar incidents, and sends a pre-populated summary to the responder.
• On-call coordination: Query who's currently on-call across multiple services, check schedule conflicts, and let Claude suggest load-balancing changes based on response time trends.

Full install guides for Claude Desktop, Cursor, Windsurf, and more at CuratedMCP.