Remote MCP server

Your monitoring data, ready for AI agents.

Connect Cursor, Codex, or Claude Code to PingView in minutes. Ask about monitors, fleet summary, latency, SSL validity, and incidents without copying data between tools.

Create API keyREST API docs
Streamable HTTP endpoint
https://pingview.app/api/mcp
user

Show my monitors

Cursor / Codex / Claude Code→list_monitors
1 monitor found
{
  "monitors": [
    {
      "name": "Storefront",
      "status": "operational"
    }
  ],
  "total": 1,
  "pageSize": 50,
  "hasMore": false
}

Production-ready safety

Focused tools expose monitoring context, metrics, and deployment maintenance windows without exposing destructive actions.

Your existing access rules stay in force

Bearer authentication, team ownership checks, and API rate limits are reused from the public API.

Connect your client

Create a PingView API key, put it in an environment variable, then register the remote endpoint.

  1. 1. Create an API key in PingView.

    Create API key

  2. 2. Set the key for your current terminal session.

    Windows PowerShell
    $env:PINGVIEW_API_KEY = "pk_live_..."
    macOS / Linux
    export PINGVIEW_API_KEY="pk_live_..."

  3. 3. MCP client

    Codex

    Codex stores the environment variable name, not the token value.

    CLI
    codex mcp add pingview --url https://pingview.app/api/mcp --bearer-token-env-var PINGVIEW_API_KEY

    Verify the connection: codex mcp get pingview

    Claude Code

    The configuration uses Claude Code environment variable expansion, so the key stays outside the saved MCP definition.

    CLI
    claude mcp add-json --scope user pingview '{"type":"http","url":"https://pingview.app/api/mcp","headers":{"Authorization":"Bearer ${PINGVIEW_API_KEY}"}}'

    Verify the connection: claude mcp get pingview

Then ask: “Use PingView and summarize my fleet.”

Comprehensive monitoring context

The agent discovers only the operations it needs to answer monitoring questions and support CI/CD deployments.

  • list_monitorsList monitors with optional status, active, and search filters.
  • get_fleet_summaryOverall fleet summary: status distribution, count, and API limits.
  • get_monitorRead one monitor configuration and its 24-hour summary.
  • get_monitor_quick_statusDense summary: status, p95 latency, trend, SSL, and current incident.
  • get_monitor_statusRead backend-computed status and per-location observations.
  • get_monitor_uptimeRead availability aggregation for 24h, 7d, 30d, or 90d.
  • get_monitor_sslRead SSL certificate validity, issuer, and days remaining.
  • get_monitor_performanceRead latency statistics (p95, min, avg, max) and TTFB.
  • list_monitor_incidentsList recent incidents for one monitor.
  • list_maintenance_windowsList active, scheduled, or past maintenance windows.
  • create_maintenance_windowSchedule a maintenance window to silence alerts during deploys.
  • cancel_maintenance_windowCancel an active maintenance window immediately.

The agent gets context, not control.

PingView remains the source of truth for status and availability. Mutating actions require an explicit write-scoped API key, and all calls are tracked in API telemetry and request logs.

Give your agent operational context.

Create a scoped API key and connect the MCP endpoint. You can revoke the key at any time.

Create API key