Model Context Protocol (MCP) Integration

UpMonitor supports the Model Context Protocol (MCP), allowing AI agents (Claude Desktop, Claude Code, Cursor, and any MCP-compatible client) to directly perform website diagnostics and monitoring tasks.

New: Active Incident handoff. When a monitor degrades, the Active Incident panel generates a copy-pasteable prompt that drops the entire incident context - root cause, remediation steps, deep link - into your AI agent in one click. The MCP server is what turns that prompt into structured, queryable context. See the agentic setup guide for the per-tool install path.

Why use MCP with UpMonitor?

By connecting UpMonitor to your AI agent, you can ask questions like:

"Is my website's SSL certificate valid?"
"Check if https://mysite.com has security headers configured."
"What's my current API usage this month?"

The agent calls the right tool, summarizes the output, and suggests fixes - all without leaving your IDE.

Configuration

Claude Desktop

Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "upmonitor": {
      "command": "npx",
      "args": ["-y", "@upmonitor/mcp-server"],
      "env": {
        "UPMONITOR_API_KEY": "up_..."
      }
    }
  }
}

Claude Code

claude mcp add upmonitor \
  -e UPMONITOR_API_KEY=up_... \
  -- npx -y @upmonitor/mcp-server

Cursor

Add to Settings → MCP → Add Server:

{
  "upmonitor": {
    "command": "npx",
    "args": ["-y", "@upmonitor/mcp-server"],
    "env": {
      "UPMONITOR_API_KEY": "up_..."
    }
  }
}

Available Tools

Audits

run_audit(url, checkers?): One-off diagnostic check.

Monitoring

list_monitors(): List your continuous monitors.
add_monitor(url, name?, interval?, checkers?): Create a monitor.
get_monitor_status(monitor_id, history?): Get state + recent history.
update_monitor(monitor_id, name?, ...): Edit an existing monitor.
delete_monitor(monitor_id): Delete a monitor.
run_check_on_monitor(monitor_id): Trigger manual check now.

Incidents

list_incidents(status?, acknowledged?, ...): View incident inbox.
get_incident(incident_id): AI diagnosis & root cause details.
get_monitor_incident(monitor_id): Check if a monitor has an active issue.
acknowledge_incident(incident_id): Silence notifications.
resolve_incident(incident_id): Close an incident manually.

Usage

get_usage(): Remaining API quota.

Security

The MCP integration uses your standard UpMonitor API key. Treat it like a password:

Never commit it to a public repository.
Rotate it via Developer Settings → API Keys, if exposed.
Use the UPMONITOR_API_KEY environment variable, not config file paths checked into git.