Ask Claude if your stack is broken.
Point any MCP client at Pingoru's public endpoint and ask whether a vendor is down, which components are affected, or what broke last week. No account or key needed.
https://pingoru.io/api/mcp-public/
{
"mcpServers": {
"pingoru": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://pingoru.io/api/mcp-public/"
]
}
}
}Rate-limited per IP · read-only · 8,000+ status pages · Read the docs →
What your AI assistant can ask
Once configured, your agent has these tools available. Read-only: Claude (or Cursor, etc.) cannot change your monitors or alerts through MCP.
provider_status() CatalogGet the current status, any active incidents and upcoming maintenance for a status-page vendor.
"Is GitHub down right now?"
provider_components() CatalogList a provider's components with their current status; optionally filter to non-operational only.
"Which Cloudflare components are degraded?"
component_status() CatalogPinpoint the status of a single component by name (case-insensitive substring match).
"What's the status of EC2 us-east-1?"
provider_incidents() CatalogRecent incidents and maintenance for a provider (last 1–30 days) with status, times, affected components and links, with optional filter to a single component.
"Any AWS issues this week?"
incident_detail() CatalogFull record for one incident or maintenance: status, start/resolve times, duration, affected components, links and every status update with its timestamp.
"Walk me through the May 3 Stripe outage."
search_providers() CatalogFind providers by name or slug. Use when the agent does not know the exact slug.
"Look up that DBaaS, planetscale?"
my_monitors() AccountList all of the caller's monitored services with their current rolled-up health.
"What's the health of my entire stack?"
my_active_alerts() AccountOpen incidents on the caller's monitored services right now.
"Is anything I monitor broken right now?"
my_recent_incidents() AccountResolved incidents on the caller's monitors over the last N days (max 30).
"What burned us this week?"
With a Pro key: your own monitors and alerts
A Pro API key adds account tools (your monitors, your active alerts, incidents across your stack). Create a key in the dashboard, then pick your tool below. Pro is $5/mo.
Anthropic's desktop chat app. The most polished MCP experience and our recommended starting point.
Don't have it yet? Install Claude Desktop →
{
"mcpServers": {
"pingoru": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://mcp.pingoru.io/",
"--header",
"Authorization: Bearer pk_REPLACE_WITH_YOUR_KEY"
]
}
}
}Replace pk_REPLACE_WITH_YOUR_KEY with the real key
shown when you created it. Endpoint is https://mcp.pingoru.io/.
Don't see your tool? Use the bring-your-own-tool config →
FAQ
What is MCP?
Model Context Protocol: an open standard from Anthropic that lets AI assistants connect to external data sources and tools. Pingoru's MCP endpoint plugs into any MCP-compatible client (Claude Desktop, Claude Code, Cursor, Cline, Zed, Goose, and more) to give your assistant real-time status data on 8,000+ cloud and SaaS providers.
Which providers can my assistant ask about?
The public endpoint answers about every provider Pingoru tracks, with no account. Through the Pro endpoint, catalog lookups cover the most-asked-about vendors (AWS, GitHub, Stripe, Cloudflare, Slack, Twilio and so on) plus any provider you have a monitor on, and the account tools add your own monitors and alerts.
Which AI tools work with this?
Claude Desktop and Claude Code are the most polished. Cursor, Cline, Continue.dev, Zed, and Goose all have native MCP support. ChatGPT works via a community-maintained bridge. If your tool isn't listed, any MCP-compatible client can connect: point it at the endpoint URL (Streamable HTTP), plus a Bearer token for the Pro endpoint.
What does it cost?
The public endpoint is free with no account, limited per IP to 30 requests a minute and 500 a day. The Pro endpoint, with your own monitors and alerts, is included in Pro ($5/mo): 500 calls per month across all keys. Need more for a power-user workflow? Reach out and we can grant an override.
Is my data safe?
Read-only by design: Claude (or any agent) cannot pause monitors, acknowledge incidents, or change anything through MCP. Each request is authenticated by an API key you create and revoke from your dashboard. We log call metadata (timestamp, tool name, status, duration) for 90 days for forensics, never the response bodies.
Can I rate-limit my own keys?
Three layers of limits ship out of the box: 30 requests/minute burst per key, 500 calls/month per key, and 500 calls/month per account. Each layer returns a structured error with a reset timestamp so your agent can back off cleanly. Need more headroom? Reach out to support and we can grant a per-account override.
How fresh is the data?
Pingoru reads each status page every five to ten minutes, so when GitHub posts an incident on its status page your assistant sees it through MCP within minutes.
Can I look up incidents from years ago?
No: incident lookback is capped at 30 days. For older history we point to the upstream vendor's status page. This keeps the protocol surface tight and avoids encouraging vendor-lock-in to our incident DB.
Can I write data through MCP?
Not in v1. Read-only by design. Write operations (pause monitor, acknowledge incident) raise consent and audit questions we want to think through carefully before exposing them to an agent loop.
My AI tool isn't in your list. Can I still use this?
Yes. Any MCP-compatible client connects with the endpoint URL, plus a Bearer token for the Pro endpoint. The "My tool isn't here" tab in the Pro setup section has the raw config.
Where do I report bugs?
Email [email protected].
Ready?
Sign up free, upgrade when you're ready. Pro is $5/mo and the MCP endpoint is included.
Get started free Compare plans